The real estate sector is a cornerstone of economic development, significantly contributing to GDP and employment. For financial institutions, assessing the creditworthiness of developers and buyers is crucial to mitigate credit risks and maintain financial stability. In this context, the purpose of this study is to propose a model for real estate development rating to enhance the decision-making process. To achieve this, we combined existing data and collected ones with insights from experts, along with a machine learning perspective, to create a multidimensional model, with its features and process, which takes into account companies, projects, and market data. For the proposed grid, the selected features include companies’ financial data, historical behavior, project ratios, location with proximity to amenities, mobility and accessibility, macro market data and other influencing ones. Applied to 653 different cases in Morocco between 2016 and 2025, the rating is based on the XGBoost probability of default and confronted to the effective default to be used by banks and developers.