Food security is a growing global concern. Given its pivotal role in food production, agriculture is a key to address global food security challenges. By integrating machine learning (ML) to optimize crop yields and resource management, it can significantly contribute to long-term stability and economic development. This paper focuses on enhancing crop recommendation systems using five ML algorithms: Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). Remarkably, the SVM model demonstrated significant improvements in accuracy, precision, and F1-score following hyperparameter tuning, achieving a testing accuracy of 99.32%. The optimized SVM model holds considerable potential for scalability and practical application in crop recommendation systems, promoting sustainable agriculture. Future research will focus on expanding the dataset using advanced techniques and enabling the system to select the most appropriate algorithm based on environmental factors, further enhancing its real-world applicability.