diff --git a/doc/source/whatsnew/v1.3.0.rst b/doc/source/whatsnew/v1.3.0.rst
index b36499c340fd9..cac7b9d8677b0 100644
--- a/doc/source/whatsnew/v1.3.0.rst
+++ b/doc/source/whatsnew/v1.3.0.rst
@@ -945,6 +945,7 @@ Indexing
 - Bug in :meth:`DataFrame.loc` returning :class:`MultiIndex` in wrong order if indexer has duplicates (:issue:`40978`)
 - Bug in :meth:`DataFrame.__setitem__` raising ``TypeError`` when using a str subclass as the column name with a :class:`DatetimeIndex` (:issue:`37366`)
 - Bug in :meth:`PeriodIndex.get_loc` failing to raise ``KeyError`` when given a :class:`Period` with a mismatched ``freq`` (:issue:`41670`)
+- Bug ``.loc.__getitem__`` with a :class:`UInt64Index` and negative-integer keys raising ``OverflowError`` instead of ``KeyError`` in some cases, wrapping around to positive integers in others (:issue:`41777`)
 
 Missing
 ^^^^^^^
diff --git a/pandas/_libs/index.pyx b/pandas/_libs/index.pyx
index f7cec262ca302..3351bb7cac7d6 100644
--- a/pandas/_libs/index.pyx
+++ b/pandas/_libs/index.pyx
@@ -106,7 +106,8 @@ cdef class IndexEngine:
 
         try:
             return self.mapping.get_item(val)
-        except (TypeError, ValueError):
+        except (TypeError, ValueError, OverflowError):
+            # GH#41775 OverflowError e.g. if we are uint64 and val is -1
             raise KeyError(val)
 
     cdef inline _get_loc_duplicates(self, object val):
diff --git a/pandas/core/indexes/base.py b/pandas/core/indexes/base.py
index 124903446220d..db718916d7fd7 100644
--- a/pandas/core/indexes/base.py
+++ b/pandas/core/indexes/base.py
@@ -5410,6 +5410,7 @@ def _find_common_type_compat(self, target) -> DtypeObj:
                 return np.dtype("object")
 
         dtype = find_common_type([self.dtype, target_dtype])
+
         if dtype.kind in ["i", "u"]:
             # TODO: what about reversed with self being categorical?
             if (
diff --git a/pandas/core/indexes/datetimelike.py b/pandas/core/indexes/datetimelike.py
index 5f24eb0cfaad6..3dc46f04d1d45 100644
--- a/pandas/core/indexes/datetimelike.py
+++ b/pandas/core/indexes/datetimelike.py
@@ -49,7 +49,6 @@
     TimedeltaArray,
 )
 from pandas.core.arrays.datetimelike import DatetimeLikeArrayMixin
-import pandas.core.common as com
 import pandas.core.indexes.base as ibase
 from pandas.core.indexes.base import (
     Index,
@@ -599,7 +598,7 @@ def _convert_arr_indexer(self, keyarr):
         try:
             return self._data._validate_listlike(keyarr, allow_object=True)
         except (ValueError, TypeError):
-            return com.asarray_tuplesafe(keyarr)
+            return super()._convert_arr_indexer(keyarr)
 
 
 class DatetimeTimedeltaMixin(DatetimeIndexOpsMixin):
diff --git a/pandas/core/indexes/numeric.py b/pandas/core/indexes/numeric.py
index de7c522b4fbec..e6526bd0eaf2f 100644
--- a/pandas/core/indexes/numeric.py
+++ b/pandas/core/indexes/numeric.py
@@ -37,7 +37,6 @@
 )
 from pandas.core.dtypes.generic import ABCSeries
 
-import pandas.core.common as com
 from pandas.core.indexes.base import (
     Index,
     maybe_extract_name,
@@ -250,21 +249,6 @@ def _maybe_cast_slice_bound(self, label, side: str, kind=lib.no_default):
         # we will try to coerce to integers
         return self._maybe_cast_indexer(label)
 
-    @doc(Index._convert_arr_indexer)
-    def _convert_arr_indexer(self, keyarr) -> np.ndarray:
-        if not is_unsigned_integer_dtype(self.dtype):
-            return super()._convert_arr_indexer(keyarr)
-
-        # Cast the indexer to uint64 if possible so that the values returned
-        # from indexing are also uint64.
-        dtype = None
-        if is_integer_dtype(keyarr) or (
-            lib.infer_dtype(keyarr, skipna=False) == "integer"
-        ):
-            dtype = np.dtype(np.uint64)
-
-        return com.asarray_tuplesafe(keyarr, dtype=dtype)
-
     # ----------------------------------------------------------------
 
     @doc(Index._shallow_copy)
diff --git a/pandas/tests/indexing/test_loc.py b/pandas/tests/indexing/test_loc.py
index ab868a3d3713d..dcccd42c52c8c 100644
--- a/pandas/tests/indexing/test_loc.py
+++ b/pandas/tests/indexing/test_loc.py
@@ -1010,18 +1010,32 @@ def test_loc_copy_vs_view(self):
     def test_loc_uint64(self):
         # GH20722
         # Test whether loc accept uint64 max value as index.
-        s = Series([1, 2], index=[np.iinfo("uint64").max - 1, np.iinfo("uint64").max])
+        umax = np.iinfo("uint64").max
+        ser = Series([1, 2], index=[umax - 1, umax])
 
-        result = s.loc[np.iinfo("uint64").max - 1]
-        expected = s.iloc[0]
+        result = ser.loc[umax - 1]
+        expected = ser.iloc[0]
         assert result == expected
 
-        result = s.loc[[np.iinfo("uint64").max - 1]]
-        expected = s.iloc[[0]]
+        result = ser.loc[[umax - 1]]
+        expected = ser.iloc[[0]]
         tm.assert_series_equal(result, expected)
 
-        result = s.loc[[np.iinfo("uint64").max - 1, np.iinfo("uint64").max]]
-        tm.assert_series_equal(result, s)
+        result = ser.loc[[umax - 1, umax]]
+        tm.assert_series_equal(result, ser)
+
+    def test_loc_uint64_disallow_negative(self):
+        # GH#41775
+        umax = np.iinfo("uint64").max
+        ser = Series([1, 2], index=[umax - 1, umax])
+
+        with pytest.raises(KeyError, match="-1"):
+            # don't wrap around
+            ser.loc[-1]
+
+        with pytest.raises(KeyError, match="-1"):
+            # don't wrap around
+            ser.loc[[-1]]
 
     def test_loc_setitem_empty_append_expands_rows(self):
         # GH6173, various appends to an empty dataframe