Phishing websites, masquerading as legitimate platforms, continue to spoof users into reveal sensitive information. In response, we introduce a novel model designed to detect fraudulent URLs, a critical step in combating phishing attacks. By implementing of broader range of deep neural network techniques and machine learning (ML) models, such as Decision Tree (DT), Support Vector Machine (SVM), XGBoost, Multilayer Perceptrons, and Autoencoder Neural Networks (ANN), we proffer a meticulous practice for specifically recognizing phishing experiments. By inventively combining these methods, version of our model not only upgrades detection accuracy but also marks a remarkable growth in the domain. After an exhaustive analysis, we use XGBoost to train our blueprint. This model has the highest accuracy rate, which helps indicate the productiveness of our approach and yields salient latest data for the ongoing conflict in opposition to online scam.