<p>Agriculture holds profound significance in the lives and livelihoods of Bangladesh’s population. The escalating population has contributed to a reduction in available arable land, exacerbating concerns about the feasibility of farming. In various regions of Bangladesh, there is a prevailing perception that certain areas are unsuitable for cultivation. Consequently, a substantial amount of land still needs to be tapped and explored for agricultural purposes, contributing to underutilization and hindering potential agricultural development in these regions. Considering these issues, this study seeks to forecast the optimal crop choices for specific soil classes, enabling individuals to make more informed decisions about crop cultivation on their land. Initially, the soil class is identified based on its unique characteristics within a given area, and subsequently, crop selection is determined based on these distinct soil classes. This study explores the performance of Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). It proposes a hybrid PSO-GA technique to optimize the Decision Tree model for forecasting crop cultivation based on classifying soil. The performance of a standalone decision tree model is also measured. In soil classification, the hybrid PSO-GA approach demonstrates superior performance, achieving an accuracy of 96.04%, surpassing individual PSO, GA, and standalone decision tree methods. In crop cultivation prediction, the hybrid approach outperforms individual PSO, GA, and decision tree methods with an accuracy of 92.63%. The results highlight the efficacy of the integrated PSO-GA strategy in optimizing the DT model for precise agricultural predictions. This research contributes valuable insights for enhancing decision support systems in agriculture, providing a promising avenue for improved accuracy in soil classification and crop cultivation prediction.</p>