Accurate determination of subsurface properties of petroleum reservoirs, particularly porosity, Young’s modulus, and reservoir pressure, plays an important role in reservoir evaluation, well design, and engineering decision-making. The primary objective of this study is to develop an advanced machine learning framework for the simultaneous estimation of these properties by combining well-logging data, mud logging unit records, and geomechanical parameters. The study population comprises data obtained from drilling operations and well-logging in oil fields. For modeling, a set of machine learning algorithms including Support Vector Machine, Gradient Boosting, AdaBoost, CatBoost, Random Forest, XGBoost, LightGBM, Extra Trees, KNN, and Regression Tree were employed, and the data, after applying preprocessing procedures, were used for training and evaluating the models. The preprocessing steps included data normalization and the removal of noise and outliers to improve model performance. Evaluation metrics included the coefficient of determination (R²), MSE, RMSE, MAE, AARE, and AAPE. The study’s findings indicate that the Gradient Boosting model achieved the best performance in estimating all three parameters, porosity, Young’s modulus, and reservoir pressure, and, in addition to high accuracy on the training data, demonstrated satisfactory generalizability to the test data. Other algorithms yielded varying results depending on the target parameter; for example, KNN and Regression Tree performed well on the training data but showed limitations in generalizing to new data. After model comparison, sensitivity analysis and Williams plots were performed for the best model. The conclusion of this research suggests that the concurrent use of well-logging data and mud logging unit records together with advanced machine learning algorithms enables accurate and reliable estimation of reservoir properties. This framework can help reduce operational costs, accelerate engineering