<p>The Caspian Sea has experienced significant interannual and seasonal fluctuations in water level over the past decades, with considerable consequences for coastal ecosystems, infrastructure, and regional economies. Understanding the drivers of these variations is crucial for effective forecasting and long-term planning. This study focuses on the numerical modeling of sea level variability in the Caspian Sea during the period 1940–2024, using atmospheric forcing from the ERA5 reanalysis. The model accounts for river runoff, visible evaporation from the sea surface, and precipitation over the sea, with verification performed against observational data from key stations: Tyuleniy Island, Makhachkala, and Fort Shevchenko. Analysis reveals that the sea level responds to the balance between river inflow and evaporation, with periods of sea level rise (1978–1995) and decline (1940–1977, 1996–2024) well captured by the model. Seasonal changes in sea level, primarily driven by Volga discharge and evaporation, are generally 20–40&#xa0;cm and are well reproduced by the model. The validation results show high correlation coefficients (0.95–0.98) and low biases, confirming the model's ability to replicate observed variations at seasonal and interannual scales. Minor discrepancies are attributed to uncertainties in river discharge and evaporation estimates. Overall, the modeling framework demonstrates strong reliability for assessing the Caspian Sea's level variability. The results demonstrate good agreement between simulated and observed sea levels, confirming the model's ability to reproduce both long-term trends and seasonal dynamics. The approach provides a reliable tool for assessing historical changes in Caspian Sea level and supports future forecasting efforts. Our results support the use of ERA5 as a reliable and consistent data source for long-term modeling of sea level variability in the Caspian Sea region.</p>