The Social Portfolio Problem (SPP) addresses the challenge of selecting the most appropriate subset of social projects under constraints of budget, time, and impact. In the context of Guadalajara, one of Mexico’s most prominent urban centers in smart city development, there is a growing need for an optimization-based decision support model that evaluates and prioritizes circular economy oriented social initiatives. This paper presents a multi-criteria selection framework integrating mathematical modeling and evolutionary algorithms namely Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) to identify an optimal project portfolio that maximizes social benefit, minimizes costs, and enhances environmental impact within a 12–18-month implementation period. Over 100 real world project proposals in Guadalajara were codified and analyzed using criteria such as population reach, cost, duration, useful life, inter-project synergy, and relevance to recycling and electronic waste (RAEE). The results indicate that the GA achieved faster convergence, while PSO provided a more diverse solution set. The proposed model supports transparent, data-driven decision-making for allocating public funds in sustainable urban development.