<p>Using machine learning (ML) techniques to study stability evaluation methods for specific types of slopes requires lots of case samples with various influence factors, obtaining sufficient reliable samples has become a challenge. For the bedding rock slope with weak interlayers (BRSWI), the progressive failure mechanism is crucial to identifying the primary controlling factors of slope stability. This study proposed a novel framework to rapidly assess slope stability in the BRSWI. Firstly, a mechanical equilibrium differential equation was driven to reveal the progressive failure mechanism of BRSWI. The strain-softening characteristics of weak interlayers were used to identify the primary controlling factors (unit weight and thickness of potential sliding block, dip angle of weak interlayer, cohesion and internal friction angle, water level, slope angle, and peak ground acceleration) of slope stability. Secondly, based on five engineering cases, numerous (163) slope samples were generated by numerical simulation under different conditions. Finally, the slope stability evaluation model was trained and selected through three ML methods (Convolutional Neural Networks—CNN, Multi-layered Perceptron—MLP, and Random Forest—RF), which used eight primary controlling factors and stability as input and output, respectively. The MLP model is recommended because the accuracy is highest (88%). The proposed method improves the efficiency of BRSWI’s stability analysis and offers potential for risk mitigation and decision support in slope management.</p>