This study presents an explainable artificial intelligence framework for climate-responsive residential envelope design in Gulf coastal cities by integrating building performance simulation, multi-objective optimization, machine learning, and explainability analysis. While previous studies have largely focused on minimizing energy consumption, limited research has simultaneously considered energy performance, capital cost, and thermal comfort within a unified and interpretable decision-support framework. The objective of this research was to identify dominant envelope design variables and derive practical design recommendations for residential buildings located in Dubai, Doha, and Manama. A two-story detached villa prototype was developed and simulated under representative coastal hot-arid climate conditions. Six envelope and operational design variables, including window-to-wall ratio (WWR), shading depth, cooling setpoint, glazing type, wall construction, and roof construction, were evaluated through a simulation-based optimization framework. A total of 600 design alternatives were generated using NSGA-II optimization and subsequently used to train Random Forest predictive models for energy use intensity (EUI), capital cost, and ASHRAE 55 thermal discomfort hours. SHAP (Shapley Additive Explanations) analysis was then applied to quantify variable importance and extract interpretable design rules. The results demonstrated strong predictive capability, with Random Forest models achieving R² values of 0.933 for EUI, 0.982 for capital cost, and 0.955 for thermal discomfort. SHAP analysis revealed that WWR was the dominant driver of energy performance, accounting for 65.2% of total feature importance, while wall construction exerted the greatest influence on capital cost. Thermal comfort was primarily governed by cooling setpoint, followed by WWR and shading depth. Dependence analysis further identified clear threshold relationships between envelope variables and performance outcomes. The proposed framework transforms optimization datasets into actionable design knowledge and provides interpretable decision support for architects, consultants, and developers seeking cost-effective and climate-responsive residential envelope solutions in Gulf coastal environments.