This paper proposes a control strategy for the energy saving of refrigerating chambers. Combining binary coding and proteome reorganization, the binary proteome algorithm (BPA) is proposed to solve this problem. The refrigeration system model is firstly established based on the performance data of compressors and temperature measurements of each refrigerating chamber. The objective function is an averaged coefficient of performance (COP), which considers the switching loss of the compressors, power consumption of the compressors, and refrigerating capacity of the chambers. The control strategy is defined as an optimization problem with constraints to avoid the ineffective operation of a refrigeration system for improving the COP. BPA is adopted to solve the control strategy for optimizing energy saving. The effectiveness and efficiency of the BPA are demonstrated using a real system, and the results are compared with the original control strategy. Results show that the average power consumption drops from 115.92 kW to 108.82 kW, and the average COP value rises from 1.92 to 2.03. The proposed control strategy is feasible, robust, and more effective in energy-saving problems. Other than energy saving, the proposed control strategy also has the benefits of reducing the evaporator frost formation, which allows the products to avoid chill damage.