Abstarct: Today, one of the main problems in the oil well drilling industry is the instability of oil wells, which increases the costs associated with drilling, operation, and problems of repair operations, as well as possible damage to production formations and oil and gas reservoirs. One of the main influential factors in maintaining and controlling the stability of the oil well wall is determining the safety window of drilling mud. The first step in well design is the mud weight design. The difference between success and failure in drilling depends on the mud weight program, which can be controlled by accurately determining the mud safe window. Uncorrected spectral gamma ray (SGR), potassium (POTA), thorium (THOR), uranium (URAN), photoelectric absorption coefficient (PEF), neutron porosity (NPHI), rock density (RHOB), corrected gamma ray (CGR) , shear wave velocity (Vs) and pressure wave velocity (Vp) are the well survey charts that are used in this research. Four random forest algorithms, support vector regression algorithm, artificial neural network algorithm and decision tree algorithm were used for evaluation. The results showed that the random forest algorithm had the highest accuracy in determining the safe window of drilling mud among the four evaluated models. In the evaluation of random forest algorithm, the pore pressure prediction accuracy value of RMSE=12.76 and R2=0.9948 and the fracture pressure prediction accuracy value of RMSE=15.71 psi, R2=0.9967 psi obtained from the side of this model show that after Training with multi-well data can also be used for other wells in the point field.