This study focuses on the development of continue to grow, strengthening cybersecurity measures a Machine Learning Algorithms and Hybrid Model is essential in protecting businesses, organizations, and that combines Naïve Bayes and all other algorithms individuals from online fraud. By improving phishing which are the XGBoost, Support Vector Machine, detection models, this research contributes to the Random Forest and Decision Tree to enhance the development of more reliable an d secure digital detection of phishing w ebsites. The system was infrastructures, ensuring safer online transactions and trained and tested using three datasets (Dataset1, communications. Dataset2 and Dataset3) with various train - test split (50% - 50%, 60% - 40%, 70% - 30%, 80% - 20%, and Meanwhile, SDG 16 aims to promote peace, justice, 90% - 10%) to evaluate its performance under and strong institutions, which includes ensuring security in both different conditions. The preprocessing stag e physical and digital spaces. Phishing is a form of cybercrime that included normalization, feature selection, and data targets individuals and organizations by tricking them into balancing using SMOTE to improve accuracy. To giving away sensitive information. This study helps in the fight validate the model, 10 - Fold Cross Validation was against cybercrime by providing effective methods to detect and applied to ensure consistency and prevent prevent phishing attacks. Strengthening cybersecurity systems overfitting. The system's performance was assessed supports the protection of personal data, financial assets, and using key eval uation metrics such as accuracy, institu tional security, ultimately contributing to a safer and more precision, recall, and F1 - score. The results revealed trustworthy digital environment [2]. the XGBoost and hybrid model is Naïve Bayes and Decision Tree classifiers, demonstrating higher detection accuracy and reliability. II. R EVIEW OF RELATED LITERATURE