This study introduces a comprehensive maintenance optimization framework for hybrid solar-wind systems, by integrating reliability-based preventive maintenance with predictive strategies based on remaining useful life (RUL) estimations. The proposed approach tackles the critical issue of minimizing maintenance costs while upholding system reliability, a key concern in renewable energy systems prone to performance losses due to component failures. The methodology leverages the Cox Proportional Hazards model to quantify reliability degradation across multiple covariates and employs Artificial Neural Networks (ANN) to forecast the RUL of essential components. A Bayesian averaging mechanism is adopted to determine optimal predictive maintenance timing, while reliability thresholds are used to initiate preventive actions. Furthermore, the model incorporates a dynamic cost-sensitivity analysis, enabling adaptive maintenance planning in response to shifting economic conditions. By analysing degradation dynamics, system topology, and cost trade-offs among preventive, predictive, and corrective maintenance, the model supports intelligent scheduling decisions. This adaptive framework enhances maintenance responsiveness to cost variations and provides a scalable solution for improving the operational efficiency of hybrid renewable systems. Ultimately, the proposed strategy serves as a robust decision-support tool for energy operators, promoting cost-effective and reliable long-term operation of hybrid solar-wind installations. The findings contribute to sustainable energy management by extending system lifespan and reducing total operational expenditure.