With the expansion of urbanization, freight transport leads to increasingly serious traffic and environmental problems. Transport planners need appropriate urban freight modeling tools to assess the impact of policy decisions on the freight transport system. This paper is designed to model the urban freight transport microscopically through three steps: freight generation, freight distribution, and vehicle routing. A heuristic optimization method that combines priority labeling and simulated annealing algorithm has been developed. The model has been applied to a real-life case study of last-mile deliveries in food retail in Berlin, which is based on existing nationwide survey data and other open source data. The final tour planning results showed that the modeled share of various types of vehicles are consistent with the real world and reflect the current state of urban freight.