This paper presents a comprehensive performance evaluation of voltage stability indices (VSIs) within a stochastic multi-objective optimal power flow (MOOPF) framework for hybrid AC/DC power systems incorporating voltage source converter–based multi-terminal DC (VSC-MTDC) transmission and high penetration of renewable energy sources (RESs). Six widely used VSIs—VCPI, LVSI, Lmn, FVSI, LQP, and NLSI—are systematically embedded as stability objectives and assessed under both normal and line-outage contingency conditions on a modified IEEE 30-bus hybrid AC/DC test system. To effectively address the nonlinear, nonconvex, stochastic, and constrained nature of the problem, a CNN-assisted cooperative bio-inspired swarm intelligent optimization framework is developed, combining deep learning–based preference guidance with competitive multi-objective swarm search. Simulation results reveal distinct operational trade-offs among the considered indices. Under normal operating conditions, the VCPI-based formulation achieves the most balanced performance, reducing generation cost and active power losses by up to 4.8% and 6.3%, respectively, while maintaining secure voltage margins. During contingency scenarios, the NLSI-based formulation demonstrates superior robustness, enhancing minimum voltage stability margins by approximately 9–10%. Furthermore, voltage-stability-constrained optimization significantly improves renewable utilization, with the VCPI case attaining the highest average RES penetration of nearly 16% of total generation. Comparative analysis confirms that the proposed CNN-assisted cooperative framework outperforms state-of-the-art multi-objective optimization techniques, providing an effective and scalable solution for secure, resilient, and sustainable operation of renewable-integrated hybrid AC/DC power systems.