This passage presents a study that addresses the issue of carbon emissions in electric vehicle cold chain logistics delivery within the context of the “dual carbon” strategic goals and the rapid growth of the fresh e-commerce industry. The research begins by utilizing polynomial functions to model vehicle speed and calculate energy consumption during delivery, which is then integrated into the cost function to determine electric vehicle power consumption costs. Given limited research on optimizing total costs and customer satisfaction as model objectives in time-varying road networks for cold chain logistics delivery, this paper develops a multi-objective electric vehicle routing optimization model. The optimization objectives include lowest total delivery cost and highest customer satisfaction. Total delivery costs encompass fixed costs for electric refrigerated vehicles, power consumption costs, refrigeration costs, carbon emission costs, and perishable product loss costs. Additionally, this paper introduces a non-dominated sorting genetic algorithm with an elite strategy (NSGA-II) to solve the proposed model. Finally, numerical experiments using data from a specific distribution network are conducted to validate both the model and solution algorithm’s effectiveness with an aim to provide decision support for cold chain logistics companies regarding reasonable deliveries.