This study aims to optimize the process of minimum quantity lubrication (MQL) grinding for Inconel 625 (IN 625) while enhancing its comprehensibility. The research employed experimental methods based on the Box-Behnken design to investigate critical parameters, including tangential force, surface roughness, specific energy, and apparent coefficient of friction. Further, machine learning techniques, specifically Random forest regression and Gaussian process regression (GPR), have been employed to build predictive models. These models have been assessed using metrics like R2, mean absolute error, and root mean square error. The results demonstrate that GPR outperforms other techniques in predicting the data accurately. Additionally, this study utilized multi-criteria decision-making techniques, namely TOPSIS and VIKOR, in conjunction with the entropy method to determine the optimal conditions for MQL grinding of IN 625. The optimized parameters for achieving low tangential force, high surface roughness, low specific energy, and low apparent coefficient of friction have been identified as a wheel speed of 1800 m/min, table speed of 9000 mm/min, and a depth of cut of 0.01 mm. Furthermore, a higher value of Ra (surface roughness) indicates the superior effectiveness of MQL grinding compared to dry grinding, as applying the MQL technique helps retain the sharpness of grit for a longer period of time. SEM image and EDS analysis of the ground surfaces confirm that better surface morphology has been obtained at optimized parameters of MQL grinding