diff --git a/tensorcircuit/templates/blocks.py b/tensorcircuit/templates/blocks.py
index 1e376903..b4983dc6 100644
--- a/tensorcircuit/templates/blocks.py
+++ b/tensorcircuit/templates/blocks.py
@@ -181,10 +181,10 @@ def qft(
             for i in range(len(index) // 2):
                 c.swap(index[i], index[len(index) - 1 - i])
         for i in range(len(index) - 1, -1, -1):
-            rotation = -np.pi / 2
+            rotation = -np.pi / 2 ** (len(index) - i - 1)
             for j in range(len(index) - 1, i, -1):
                 c.cphase(index[j], index[i], theta=rotation)
-                rotation /= 2
+                rotation *= 2
             c.H(index[i])
             if insert_barriers:
                 c.barrier_instruction(range(min(index), max(index) + 1))
diff --git a/tests/test_templates.py b/tests/test_templates.py
index 79e2b176..219f64f0 100644
--- a/tests/test_templates.py
+++ b/tests/test_templates.py
@@ -66,10 +66,20 @@ def test_bell_block():
 
 
 def test_qft_block() -> None:
-    c = tc.Circuit(4)
-    c = tc.templates.blocks.qft(c, 0, 1, 2, 3)
-    s = c.perfect_sampling()
-    assert s[1] - 0.0624999 < 10e-6
+    n_qubits = 4
+    c = tc.Circuit(n_qubits)
+    c = tc.templates.blocks.qft(c, *range(n_qubits))
+    mat = c.quoperator().eval().reshape(2 ** (n_qubits), -1)
+    N = 2**n_qubits
+    ref = np.exp(
+        1j * 2 * np.pi * np.arange(N).reshape(-1, 1) * np.arange(N).reshape(1, -1) / N
+    ) / np.sqrt(N)
+    np.testing.assert_allclose(mat, ref, atol=1e-7)
+
+    c = tc.Circuit(n_qubits)
+    c = tc.templates.blocks.qft(c, *range(n_qubits), inverse=True)
+    mat = c.quoperator().eval().reshape(2 ** (n_qubits), -1)
+    np.testing.assert_allclose(mat, ref.T.conj(), atol=1e-7)
 
 
 def test_grid_coord():