diff --git a/CHANGELOG.md b/CHANGELOG.md
index 9f5ba02196..588148b761 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -29,7 +29,8 @@ but cannot always guarantee backwards compatibility. Changes that may **break co
 - Fixed a bug when loading the weights of a `TorchForecastingModel` trained with encoders or a Likelihood. [#1744](https://github.com/unit8co/darts/pull/1744) by [Antoine Madrona](https://github.com/madtoinou).
 - Fixed a bug when using selected `target_components` with `ShapExplainer. [#1803](https://github.com/unit8co/darts/pull/#1803) by [Dennis Bader](https://github.com/dennisbader).
 - Fixed `TimeSeries.__getitem__()` for series with a RangeIndex with start != 0 and freq != 1. [#1868](https://github.com/unit8co/darts/pull/#1868) by [Dennis Bader](https://github.com/dennisbader).
-- Fixed an issue where `DTWAlignment.plot_alignment` was not plotting the alignment plot of series with a RangeIndex correctly. [#1880](https://github.com/unit8co/darts/pull/1880) by [Ahmet Zamanis](https://github.com/AhmetZamanis) and [Dennis Bader](https://github.com/dennisbader).
+- Fixed an issue where `DTWAlignment.plot_alignment()` was not plotting the alignment plot of series with a RangeIndex correctly. [#1880](https://github.com/unit8co/darts/pull/1880) by [Ahmet Zamanis](https://github.com/AhmetZamanis) and [Dennis Bader](https://github.com/dennisbader).
+- Fixed an issue when calling `ARIMA.predict()` and `num_samples > 1` (probabilistic forecasting), where the start point of the simulation was not anchored to the end of the target series. [#1893](https://github.com/unit8co/darts/pull/1893) by [Dennis Bader](https://github.com/dennisbader).
 
 **Removed**
 - Removed support for Python 3.7 [#1864](https://github.com/unit8co/darts/pull/#1864) by [Dennis Bader](https://github.com/dennisbader).
diff --git a/darts/models/forecasting/arima.py b/darts/models/forecasting/arima.py
index 33ca7c2efe..e3149aa2fc 100644
--- a/darts/models/forecasting/arima.py
+++ b/darts/models/forecasting/arima.py
@@ -161,6 +161,7 @@ def _predict(
                 repetitions=num_samples,
                 initial_state=self.model.states.predicted[-1, :],
                 random_state=self._random_state,
+                anchor="end",
                 exog=future_covariates.values(copy=False)
                 if future_covariates
                 else None,