Abstract: The selection of the last-mile delivery mode for fresh e-commerce has a direct impact on customer experience and enterprise operational efficiency. We innovatively integrate dual perspectives from both customers and enterprises, constructing a utility maximization model for bounded rational customers and a cost minimization model for fully rational enterprises. Through these models, this paper provides an in-depth analysis of the differences in distribution mode choice behavior and their effects on service quality and enterprise costs. From the customer perspective, considering the limitations of customer rationality, this paper develops a model aimed at maximizing the average expected utility of distribution services to explore the behavioral logic of customers in choosing distribution modes. From the enterprise perspective, under the carbon tax policy, a timedependent vehicle routing optimization model is employed to optimize the calculation of different last-mile delivery modes, thereby determining the minimum cost for each mode. The results indicate that customer rationality, pickup distance, and distribution costs significantly influence utility perception. Enterprises can reduce distribution costs through time-dependent vehicle routing optimization under carbon tax policies. A case study demonstrates that ignoring customer bounded rationality leads to deviations in enterprise utility estimation. These research findings provide a theoretical foundation for the design and policy formulation of optimized delivery services for fresh e-commerce.