Groundwater is one of the most critical sources of water supply, particularly in arid and semi-arid regions. Accurate simulation of groundwater levels is challenging due to the complexity of aquifer systems and inherent uncertainties. This study evaluates three intelligent models, decision tree, random forest, and fuzzy inference system for simulating groundwater levels in the Talesh aquifer during the period 2002–2015. Input parameters included precipitation, temperature, groundwater extraction, evaporation, and groundwater level from the previous month. Seventy percent of the data was used for training the models, while the remaining 30% was utilized for testing. The results indicated that the combination of groundwater level from the previous month, precipitation, groundwater extraction, and evaporation (Pattern C) provided the most accurate simulation. Among the input parameters, groundwater level from the previous month was identified as the most influential factor. The random forest model outperformed the others, achieving RMSE and MAE values of 0.44 m and 0.35 m, respectively. In contrast, the fuzzy inference system exhibited the lowest accuracy, with RMSE and MAE values of 0.49 m and 0.389 m, respectively.