diff --git a/pandas/core/dtypes/astype.py b/pandas/core/dtypes/astype.py
index 1e78bf0cd33ae..7636cd9080107 100644
--- a/pandas/core/dtypes/astype.py
+++ b/pandas/core/dtypes/astype.py
@@ -224,7 +224,7 @@ def astype_array(values: ArrayLike, dtype: DtypeObj, copy: bool = False) -> Arra
         values = values.astype(dtype, copy=copy)
 
     else:
-        values = astype_nansafe(values, dtype, copy=copy)
+        values = astype_nansafe(values, dtype, copy=copy, skipna=True)
 
     # in pandas we don't store numpy str dtypes, so convert to object
     if isinstance(dtype, np.dtype) and issubclass(values.dtype.type, str):
diff --git a/pandas/tests/dtypes/test_dtypes.py b/pandas/tests/dtypes/test_dtypes.py
index f077317e7ebbe..ecde19ab48911 100644
--- a/pandas/tests/dtypes/test_dtypes.py
+++ b/pandas/tests/dtypes/test_dtypes.py
@@ -1107,3 +1107,15 @@ def test_compare_complex_dtypes():
 
     with pytest.raises(TypeError, match=msg):
         df.lt(df.astype(object))
+
+
+@pytest.mark.parametrize(
+    "array",
+    [(["foo", "bar", pd.NA]), (["foo", "bar", np.NaN]), (["foo", "bar", None])],
+)
+def test_skip_convert_nan_values_to_strings(array):
+    # GH 44156
+    result = Series(array).astype("str")
+    expected = Series(array)
+
+    tm.assert_series_equal(result, expected)