diff --git a/src/TiledArray/tensor/print.h b/src/TiledArray/tensor/print.h
index a2497bff15..18d09be7a7 100644
--- a/src/TiledArray/tensor/print.h
+++ b/src/TiledArray/tensor/print.h
@@ -40,12 +40,45 @@ namespace detail {
 class NDArrayPrinter {
  public:
   NDArrayPrinter(int width = 10, int precision = 6)
-      : width(width), precision(precision) {}
+      : width(width),
+        precision(precision),
+        truncate_(0.5 * std::pow(10., -precision)) {}
 
  private:
   int width = 10;
   int precision = 10;
 
+  /// truncates (=sets to zero) small floating-point numbers
+  class FloatTruncate {
+   public:
+    /// truncates numbers smaller than @p threshold
+    FloatTruncate(double threshold) noexcept : threshold_{threshold} {}
+
+    [[nodiscard]] auto operator()(std::floating_point auto val) const noexcept {
+      return std::abs(val) < threshold_ ? decltype(val){0} : val;
+    }
+
+    template <typename T>
+      requires detail::is_complex_v<T> &&
+               std::floating_point<typename T::value_type>
+    [[nodiscard]] auto operator()(T const& val) const noexcept {
+      using std::imag;
+      using std::real;
+      return T{(*this)(real(val)), (*this)(imag(val))};
+    }
+
+    template <typename T>
+      requires(!(std::floating_point<T> || detail::is_complex_v<T>))
+    [[nodiscard]] auto operator()(T const& val) const noexcept {
+      return val;
+    }
+
+   private:
+    double threshold_;
+  };
+
+  FloatTruncate truncate_;
+
   // Helper function to recursively print the array
   template <typename T, typename Index = Range1::index1_type,
             typename Char = char, typename CharTraits = std::char_traits<Char>>
diff --git a/src/TiledArray/tensor/print.ipp b/src/TiledArray/tensor/print.ipp
index 8634418138..c3535409b3 100644
--- a/src/TiledArray/tensor/print.ipp
+++ b/src/TiledArray/tensor/print.ipp
@@ -50,9 +50,9 @@ void NDArrayPrinter::printArray(const T* data, const std::size_t order,
 
   for (size_t i = 0; i < extents[level]; ++i) {
     if (level == order - 1) {
+      auto value = truncate_(data[offset + i * strides[level]]);
       // At the deepest level, print the actual values
-      os << std::fixed << std::setprecision(precision) << std::setw(width) << std::setfill(Char(' '))
-         << data[offset + i * strides[level]];
+      os << std::fixed << std::setprecision(precision) << std::setw(width) << std::setfill(Char(' ')) << value;
       if (i < extents[level] - 1) {
         os << ", ";
       }