This study quantitatively evaluates CO2 emissions from the transportation and delivery stages of urban food retail operations, aiming to identify store-level logistical characteristics and structural challenges that influence emissions, and to explore potential strategies for building a more sustainable food system. Although transportation and delivery account for less than 1% of total emissions in the food supply chain, significant variation exists across retail stores due to differences in logistics structures and delivery efficiency. Using the fuel efficiency method outlined in government guidelines, we calculated annual emissions for 88 stores operated by a railway-affiliated supermarket chain in the Tokyo–Yokohama metropolitan area. A case study comparing store A and S revealed how sales volume and mixed-load delivery influence emission levels. Benchmarking against Seven & I Holdings highlighted inefficiencies in delivery density and emissions per unit of sales floor area. Further analysis showed that truck size restrictions at station-adjacent stores and the centralized location of the distribution center contribute to increased delivery frequency and longer transport distances. Based on these findings, we propose several improvement strategies, including route redesign, adoption of electric and hybrid vehicles, dynamic route optimization using AI and IoT, and the development of decentralized logistics hubs. These measures, combined with enhanced traceability and integrated resource management under the food–energy–water (FEW) nexus, offer a pathway toward more sustainable and low-carbon urban food systems.