［Purposes ］ This study explores the problem of fresh cold chain center selection and routing under the conditions of differentiated customer demands and time -varying traffic.It aims to provide decision -making support for precise services targeting different customer segments，distribution center selection，and time -varying routing planning.［Methods］ To minimize the total system cost （including vehicle costs，distribution center operation costs，time window penalty costs，spoilage loss costs，fuel consumption costs，and carbon emission costs ），the model incorporated constraints reflecting the demands of different customer groups based on customer profile，time -varying speeds，and vehicle load capacities.A two -stage heuristic algorithm was developed for simulation experiments：Stage 1 allocated customers to distribution centers using a heuristic selection algorithm based on spatiotemporal distance.Stage 2 optimized routes using a genetic algorithm （GA）.［Findings］ The user profile -driven differentiation strategy significantly improves service quality，with the core customer group receiving optimal distribution services.Compared with that of the traditional genetic algorithm and the K-means clustering + genetic algorithm，the optimal cost of the two -stage heuristic algorithm is reduced by 9.15% and 12.61%，respectively.Meanwhile，the scheme under time -varying speed can better balance cost and timeliness than that under fixed speed，with the total driving time and cost within a reasonable range.［Conclusions ］ The proposed model and two -stage algorithm provide decision -making support for fresh cold chain enterprises to implement differentiated distribution strategies.