Given the importance of pan evaporation in agriculture and water resources across different climates, and considering the climatic changes in recent years, this study investigates the variations in pan evaporation and simulates it using frequency decomposition at evaporation stations located in Lorestan Province, western Iran. The results of examining the changes in pan evaporation values in the study area showed that during the cold months of the year (October to February), most stations experienced a significant decreasing trend in pan evaporation, whereas in warmer months such as March and April, an increasing trend was predominant. A hybrid Wavelet-Random Forest model was used to simulate the pan evaporation values. The evaluation results of this model using RMSE, NSE, WI, and correlation coefficient criteria indicated that this method has high accuracy in both training and testing phases. Specifically, the model was able to effectively identify seasonal and sudden changes in evaporation. Although a higher relative error was observed at some stations, such as Poldokhtar, compared to other stations, on average (across the 8 studied stations), an error rate of 2.23 mm/day, a Nash-Sutcliffe coefficient of over 0.8, and a Willmott index of over 94% were achieved in this study. Considering the 95% confidence intervals of the simulation and the provided violin plot, the performance of the hybrid model was assessed as acceptable. The overall results indicate that the combination of Wavelet transform and the Random Forest algorithm can be used as an effective tool for predicting pan evaporation.