This paper proposes an adaptive particle swarm optimization (APSO) algorithm tailored to the operational requirements of urban heating systems in complex environments. The algorithm achieves efficient optimization of heating systems under varying load conditions by adaptively adjusting inertia weights and learning factors. Firstly, a multi-objective optimization model for the heating system is established, considering energy efficiency, response time, and operational costs (OCs) as optimization indices to form a unified optimization objective function. The proposed model evaluates energy efficiency, response time, and OCs using Pareto front solutions to achieve a balanced trade-off for the heating system. Subsequently, by introducing a real-time feedback mechanism, the APSO algorithm dynamically updates its parameters based on changes in ambient temperature and load, thereby enhancing the algorithm's response speed and adaptability. Experimental results across various scenarios demonstrate that the APSO algorithm outperforms traditional PSO and rule-based control methods in terms of energy efficiency, response time, and OCs. The experimental results indicate that the APSO algorithm has significant advantages in the intelligent optimization of heating systems in complex environments and can provide reliable technical support for practical applications.