diff --git a/doc/release/upcoming_changes/19513.new_feature.rst b/doc/release/upcoming_changes/19513.new_feature.rst
new file mode 100644
index 000000000000..5f945cea2b68
--- /dev/null
+++ b/doc/release/upcoming_changes/19513.new_feature.rst
@@ -0,0 +1,4 @@
+Preliminary support for `windows/arm64` target
+----------------------------------------------
+``numpy`` added support for windows/arm64 target. Please note
+``OpenBLAS`` support is not yet available for windows/arm64 target.
diff --git a/numpy/core/include/numpy/npy_cpu.h b/numpy/core/include/numpy/npy_cpu.h
index 065176ac5fb6..bc1fad72f768 100644
--- a/numpy/core/include/numpy/npy_cpu.h
+++ b/numpy/core/include/numpy/npy_cpu.h
@@ -63,7 +63,8 @@
     #define NPY_CPU_HPPA
 #elif defined(__alpha__)
     #define NPY_CPU_ALPHA
-#elif defined(__arm__) || defined(__aarch64__)
+#elif defined(__arm__) || defined(__aarch64__) || defined(_M_ARM64)
+    /* _M_ARM64 is defined in MSVC for ARM64 compilation on Windows */
     #if defined(__ARMEB__) || defined(__AARCH64EB__)
         #if defined(__ARM_32BIT_STATE)
             #define NPY_CPU_ARMEB_AARCH32
@@ -72,10 +73,10 @@
         #else
             #define NPY_CPU_ARMEB
         #endif
-    #elif defined(__ARMEL__) || defined(__AARCH64EL__)
+    #elif defined(__ARMEL__) || defined(__AARCH64EL__) || defined(_M_ARM64)
         #if defined(__ARM_32BIT_STATE)
             #define NPY_CPU_ARMEL_AARCH32
-        #elif defined(__ARM_64BIT_STATE)
+        #elif defined(__ARM_64BIT_STATE) || defined(_M_ARM64)
             #define NPY_CPU_ARMEL_AARCH64
         #else
             #define NPY_CPU_ARMEL
diff --git a/numpy/core/tests/test_scalarprint.py b/numpy/core/tests/test_scalarprint.py
index 620472683391..2f1c3bc5e309 100644
--- a/numpy/core/tests/test_scalarprint.py
+++ b/numpy/core/tests/test_scalarprint.py
@@ -154,7 +154,8 @@ def test_dragon4(self):
                     "0.00000000000000000000000000000000000000000000140129846432"
                     "4817070923729583289916131280261941876515771757068283889791"
                     "08268586060148663818836212158203125")
-        assert_equal(fpos64(0.5**(1022 + 52), unique=False, precision=1074),
+        
+        assert_equal(fpos64(5e-324, unique=False, precision=1074),
                     "0.00000000000000000000000000000000000000000000000000000000"
                     "0000000000000000000000000000000000000000000000000000000000"
                     "0000000000000000000000000000000000000000000000000000000000"
diff --git a/numpy/random/src/philox/philox.h b/numpy/random/src/philox/philox.h
index c72424a975ef..8844acc15725 100644
--- a/numpy/random/src/philox/philox.h
+++ b/numpy/random/src/philox/philox.h
@@ -33,10 +33,16 @@ static NPY_INLINE uint64_t mulhilo64(uint64_t a, uint64_t b, uint64_t *hip) {
   return (uint64_t)product;
 }
 #else
-#ifdef _WIN32
+#if defined(_WIN32)
 #include <intrin.h>
 #if defined(_WIN64) && defined(_M_AMD64)
 #pragma intrinsic(_umul128)
+#elif defined(_WIN64) && defined(_M_ARM64)
+#pragma intrinsic(__umulh)
+static NPY_INLINE uint64_t _umul128(uint64_t a, uint64_t b, uint64_t *high) {
+  *high = __umulh(a, b);
+  return a * b;
+}
 #else
 #pragma intrinsic(__emulu)
 static NPY_INLINE uint64_t _umul128(uint64_t a, uint64_t b, uint64_t *high) {