The agricultural sector confronts significant obstacles due to crop pests and diseases, which lead to considerable damage and losses.. Timely and accurate identification of these issues is essential for managing pests effectively and securing food resources. This paper provides a machine learning based approach to detecting and categorizing crop pests and diseases, utilizing algorithms like Support Vector Machine (SVM), Random Forest, XGBoost, Logistic Regression, Improved Genetic Algorithm (IGA), and Genetic Programming. The dataset from Kaggle includes data on crop conditions, pest presence, and disease symptoms. Data preprocessing includes handling absent values, encoding categorical variables, and normalizing features. Feature engineering, including polynomial feature creation and Recursive Feature Elimination (RFE), enhances model performance by capturing complex relationships. GridSearchCV was employed for hyperparameter optimization, and Models were estimated using metrics like F1 score, recall, precision, correctness, and ROC curves. It is noticed that, XGBoost demonstrated the highest performance comparatively, attaining an accuracy of 86%, positioning it as the best performing model to identify pests and illnesses in crops. This research explains the potential of machine learning to improve agricultural sustainability and productivity, with future work set to incorporate additional data sources and advanced methods to refine accuracy and robustness.