Regular physical activity is very essential in maintaining health and fitness. But the modern work schedules too often compress time left for exercise which often turns tough and will turn toward health issues such as obesity. The exercise nutrition correlation is a critical function in fashioning the future wellness and health trends. As a response, we introduced a new hybrid machine learning model called CPE-Net, designed to predict the precise calorie consumption during physical activities. The approach for the task was given below: The CPE-Net model utilized PCA for the reduction in dimensionality and XGBoost for predictive tasks. It has been trained on more than 20,000 records with the accuracy of the prediction being 99%. Further tests through an SVR provided an accuracy of 92.3% with an RMSE of 9.2. Techniques were applied in using cross validation to ensure that the model is robust and reliable. The contribution of features for efficient calorie prediction is emphasized in the study and offers an efficient tool for personalized fitness tracking.