Agriculture plays a vital role in the global economy, directly influencing food security and livelihoods. However, the sector faces challenges such as climate variability, diverse soil conditions, and limited access to advanced farming techniques. This paper presents a machine learning (ML)-based crop classification and recommendation system designed to assist farmers in selecting the most suitable crops. The system implements a multi-model training methodology that incorporates random forest, support vector machine (SVM), XGBoost, and gradient boosting algorithms, augmented by a voting ensemble to ensure robust predictive capabilities. Feature selection is optimized through a combination of manual selection of top features and automated random forest-based feature selection. The framework’s reliability is validated through rigorous model evaluation, which includes cross-validation and test set assessment. By integrating advanced ML techniques, this study aims to support precision agriculture and promote sustainable farming practices.