Road accidents significantly contribute to traffic congestion, causing severe impacts on public health, the environment, and the economy. This study focuses on mapping road accident susceptibility in Casablanca using machine learning algorithms integrated with Geographic Information Systems (GIS). By analysing spatial relationships between various causality factors and accident pixels, models such as Random Forest (RF), Support Vector Machine (SVM), K-nearest neighbours (KNN), and Artificial Neural Network (ANN) were employed. The resulting predictive maps, particularly from the RF model with an AUC of 0.884, demonstrate high accuracy and are crucial tools for road safety management. These susceptibility maps are essential for optimizing planning, management, and safety measures on roadways, offering proactive solutions to mitigate traffic accidents in urban environments.