Numerous factors affect the thermal performance of building envelopes, and their variation levels are relatively high. The number of combination schemes is large, and the computational workload for dynamic building energy consumption simulation of all combination conditions one by one is huge, so it is worthwhile to conduct in-depth research on how to design the computational scheme to reduce the computational workload, and moreover, to ensure the validity and reliability of the computational results. It was proposed using the uniform design method to design the energy consumption simulation scheme, and its feasibility was verified. A multi-objective optimization model oriented towards building energy consumption, life cycle cost, and carbon emissions was established and solved via the NSGA-Ⅱ (non-dominated sorting genetic algorithm Ⅱ). The weights of the optimization objectives are determined using the entropy weight-TOPSIS (technique for order preference by similarity to ideal solution) method, subsequently, the optimal set of solutions was ranked. Eventually, a calculation method for determining the optimal insulation layer thickness of building envelopes was proposed. Taking a typical public building in Xuzhou as an example, the optimal thickness of the XPS (expanded polystyrene board) thermal insulation layer for exterior walls and roofs was calculated under different heating and cooling conditions in winter and summer. Results show that: compared with the limit value of thermal performance stipulated in the current energy-saving standard, the thermal performance of the exterior wall and the roof has been improved by 32.7% ~ 41.8% and 15.0% ~ 23.0%, respectively, despite the fact that the cost of the whole lifecycle has been increased by 3.6% ~ 8.1%, and the building energy consumption and whole-life carbon emissions were reduced by 4.0% ~ 9.3% and 2.2% ~ 3.7%, respectively. Good comprehensive benefits in energy conservation, economy, and environment can be given.