The inherent intermittency and stochastic nature of solar energy pose significant challenges to the stability of continuous power supply. To address these vulnerabilities, this paper provides a comprehensive review of multi-energy coupling systems (MECS) centered on solar energy. From a technical pathway perspective, three primary paradigms are systematically categorized: the PV-dominated electricity-shifting model, the CSP-centered thermal-power synergy model, and hybrid integration models involving hydrogen, geothermal, and biomass energy. Regarding optimization, the study evaluates the state-of-the-art application of Artificial Intelligence (AI) and Distributed Robust Optimization (DRO) in achieving source-load balancing across multiple time-spatial scales. Through engineering case studies, system performance is analyzed across three critical scenarios: grid stabilization in large-scale renewable bases, operational decarbonization in industrial parks, and off-grid survivability in extreme environments (e.g., high-altitude cold regions, deserts, and islands). Despite the benefits in carbon reduction and cost efficiency, bottlenecks remain in conversion efficiency and complex coordinated control. Distinct from descriptive reviews, this work transcends conventional technology enumeration by establishing a tripartite paradigm framework for solar MECS and delivering a quantitative, scenario-specific comparative analysis that reveals the boundaries of technical applicability under varying constraints. Finally, this paper proposes a future roadmap integrating digital twins and refined electricity market mechanisms to provide a theoretical foundation for resilient, next-generation energy systems.