Options trading has gained popularity among investors and academics for its flexibility and strategic potential. This study aims to enhance the performance of a bull-side put spread options trading strategy using machine learning. We integrate raw data attributes, implied volatility metrics, and technical indicators into Decision Tree and Logistic Classification models to predict trade profitability. The results reveal that while the Logistic Classification model achieves higher cumulative payoffs, it executes less reliable trades. In contrast, the Decision Tree algorithm demonstrates superior precision in identifying profitable trades, making it more effective for implementing the strategy. This research contributes to the growing literature on applying machine learning in options trading and highlights the importance of considering both predictive accuracy and trade reliability. Future research could explore advanced algorithms and incorporate market sentiment analysis to further refine the models and enhance their practical utility.