In recent years, soft computing and artificial intelligence have been used increasingly in geotechnical applications. Predicting critical slip surface (CSS) and factor of safety (FOS) of finite soil slopes is essential in geotechnical engineering. This article presents machine learning (ML) techniques on 616 data points in predicting the CSS and FOS of finite slopes. Robust ML algorithms are trained to predict these factors based on data from an analytical formulation based on the Morgenstern-Price method of slices on finite soil slopes. The input parameters considered in the present study are the slope angle \(\left( \theta \right)\) , stability number (c/γH), and internal friction angle \(\left( \phi \right)\) . The outputs are the FOS, coordinates of the critical center (xcritical, ycritical), and radius of the CSS (Rcritical). Various ML models such as multivariate linear regression, support vector regression, Gaussian process regression, tree ensemble algorithm, and multilayer perceptron are employed in predicting the FOS parameter. Deep learning is employed for the prediction of CSS components. The models are evaluated using the R2 statistical metric. The best prediction capabilities are exhibited by the GPR model with train, validation, and test performance metrics of 99.9%, 99.8%, and 99.8%, respectively. Deep neural network models show promising results in predicting the xcritical, ycritical, and Rcritical values with test performance of 93.7%, 97.5%, and 99.8%, respectively. Catering to the wider geotechnical audience, the current work presents SHAP analysis to enhance the interpretability of the best-chosen ML model. Additionally, a comparison between predicted and actual FOS for various values of \(\theta\) , c/γH, and \(\phi\) is also presented. These comparison charts also point to the inherent flaw in the XGBoost model. The study concludes that the proposed ML approach effectively predicts CSS and FOS.