diff --git a/src/_pytest/python_api.py b/src/_pytest/python_api.py
index 4174a55b5..a56eb05dd 100644
--- a/src/_pytest/python_api.py
+++ b/src/_pytest/python_api.py
@@ -455,8 +455,21 @@ class ApproxScalar(ApproxBase):
         if math.isinf(abs(self.expected)):
             return False
 
-        # Return true if the two numbers are within the tolerance.
-        result: bool = abs(self.expected - actual) <= self.tolerance
+        # Return true if the two numbers are within the tolerance.  In order to
+        # be flexible about which types are supported, try making the
+        # comparison in two ways.  The first requires that the actual value can
+        # be subtracted from the expected value.  This is necessary for complex
+        # numbers, which don't implement comparison operators.  The second
+        # requires that the actual and expected values can be compared.  This
+        # is necessary for comparing Decimals with floats, see #8495.
+        result: bool
+        try:
+            result = abs(self.expected - actual) <= self.tolerance
+        except TypeError:
+            low = self.expected - self.tolerance
+            high = self.expected + self.tolerance
+            result = low <= actual <= high
+
         return result
 
     # Ignore type because of https://github.com/python/mypy/issues/4266.
diff --git a/testing/python/approx.py b/testing/python/approx.py
index 69743cdbe..59bf9ad54 100644
--- a/testing/python/approx.py
+++ b/testing/python/approx.py
@@ -556,15 +556,20 @@ class TestApprox:
 
     def test_decimal(self):
         within_1e6 = [
-            (Decimal("1.000001"), Decimal("1.0")),
-            (Decimal("-1.000001"), Decimal("-1.0")),
+            (Decimal("1.0000005"), Decimal("1.0"), Decimal),
+            (Decimal("-1.0000005"), Decimal("-1.0"), Decimal),
+            (Decimal("1.0000005"), 1.0, float),
+            (Decimal("-1.0000005"), -1.0, float),
         ]
-        for a, x in within_1e6:
+        for a, x, T in within_1e6:
+            # Need to test the default values here, because some code is needed
+            # to account for the fact that you can't add floats to decimals.
             assert a == approx(x)
-            assert a == approx(x, rel=Decimal("5e-6"), abs=0)
-            assert a != approx(x, rel=Decimal("5e-7"), abs=0)
-            assert approx(x, rel=Decimal("5e-6"), abs=0) == a
-            assert approx(x, rel=Decimal("5e-7"), abs=0) != a
+
+            assert a == approx(x, rel=T("1e-6"), abs=0)
+            assert a != approx(x, rel=T("1e-7"), abs=0)
+            assert approx(x, rel=T("1e-6"), abs=0) == a
+            assert approx(x, rel=T("1e-7"), abs=0) != a
 
     def test_fraction(self):
         within_1e6 = [