The conventional materials-based solar cells are lagging behind due to its performance limitations and stability issues. Therefore, researchers have moved their studies in the exploration of new materials-based solar cells. In this context, several simulators have been utilized but they take lot of time in calculation and simulations. To overcome this, here different machine learning (ML) models like support vector regression (SVR), random forest (RF), and extreme gradient boosting (XGB) have been applied. Firstly, 500 dataset has been obtained through the simulator by varying thickness, mobility and defect 0.1 to 1 µm, 10 to 50 cm2/V-S, and 1 × 1010 to 1 × 1014 /cm2 respectively in the SCAPS 1d simulator. The SHAP analysis further reveals the relative importance of each property in determining device performance. The best suited model is XGB among all due to its high R2 value (0.9999) and low MSE value (0.0010) whereas LR model provides lowest ML matrices. The best PCE predicted for the proposed cell is 16.87%. The results demonstrate that variations in thickness, defect density, and electron mobility have a significant impact on PCE, and the machine learning models provide accurate predictions, offering insights for optimizing the efficiency of Sb₂Se₃-based solar cells. This work will make a significant contribution to the design and fabrication of solar cells in the manufacturing industry.