Cowpea (Vigna unguiculata L. Walp.) is a climate-resilient legume essential for food security and dryland farming in semi-arid regions of India. This study delineates climate-smart suitability zones by integrating Ensemble Species Distribution Modelling (ESDM) with Analytic Hierarchy Process-Multi-Criteria Evaluation (AHP-MCE) to incorporate climatic, soil, and land-use determinants. Seven modelling algorithms (GLM, GAM, MARS, CTA, RF, ANN, SVM) were applied under baseline, 2050, and 2070 climates (RCP 4.5 and 8.5), achieving strong ensemble performance (AUC = 0.84-0.88; TSS = 0.62-0.67). Precipitation variables dominated current suitability, while temperature extremes shaped future patterns. Integrating soil and LULC data through AHP-MCE substantially improved spatial precision, expanding optimum suitability zones and identifying new target regions in central, southern and eastern India. The integrated framework offers practical value for breeding programs (identifying heat- and drought-prone target environments), agricultural policy and investment (site-specific irrigation, diversification, climate-risk planning), and farm-level decision making. By quantitatively combining climatic, soil and land-use predictors in a unified GIS-based ensemble system, an advancement over existing legume suitability studies, this work provides a scalable tool for climate-resilient crop planning and adaptive agricultural development.