Building energy-saving design needs to consider energy consumption and thermal comfort. The research aims to propose an efficient Multi-Objective Optimization (MOO) method adapted to building energy-saving design. The research first constructs an MOO model that minimizes the annual total energy consumption and maximizes indoor thermal comfort, and clarifies the decision-making elements and constraint boundaries. According to the decomposition idea of the multi-objective evolutionary algorithm, the weight vector generation, neighborhood update and mutation strategy are improved, and an improved MOO is constructed. The method is verified by residential building data in four typical climate zones in China. When this algorithm iterates for 500 generations, the Generational Distance (GD) is only 0.025, and the inverted GD reaches 0.045. The convergence and solution set distribution uniformity are significantly better than those of the comparison algorithm. The energy consumption of the solutions produced by this algorithm is 29.5%-38.6% lower than those of the comparison algorithm, with Guangzhou as low as 31.2kWh/m2·year and Harbin 42.3kWh/m2·year. Carbon emissions are reduced by 30.1%-39.8%, with Guangzhou down to 20.5kgCO2/m2·year. The thermal comfort compliance time accounts for more than 92% with Guangzhou reaching a maximum of 96.8%. The initial investment costs are reduced by 11.7%-16.4%. This algorithm is superior to existing algorithms in convergence and uniformity of solution set distribution, which can adapt to the energy-saving design across climate zones, provide practical technical support for energy-saving and cost control throughout the building’s life cycle, and has important engineering application value.
