Refrigerated transport by road is essential for the food industry but also contributes significantly to global energy consumption. In particular, last-mile transport, where the doors of the cooling chamber are opened frequently, puts a strain on energy efficiency and temperature control due to the high heat ingress from outside into the cooling chamber. These difficulties can be reduced by thermal energy storage systems, such as secondary loop refrigeration systems, if combined with a sophisticated control scheme. Although the storage capacity of such systems is critical for the overall performance of the cooling system, little research was performed regarding the sizing of the secondary loop thermal storage capacity. Therefore, this article examines the effect of the secondary loop thermal storage capacity on energy consumption and controller performance utilizing closed-loop simulations of a refrigerated vehicle model. Both a mixed-integer model predictive control scheme that can anticipate door openings and a conventional temperature controller are analyzed. An optimal thermal storage capacity of the secondary loop is found with the model predictive controller, whereas the conventional controller cannot exploit the secondary loop and thus shows significantly inferior performance. By using a dimensionless parameter for the thermal storage capacity of the secondary loop, the optimum found can be easily applied to refrigerated vehicles with various cooling chamber dimensions.