To solve the problems of significant peak load variations and power supply shortages within smart grid systems, this paper proposes an incentive scheme for peak and off-peak time-of-use (TOU) pricing based on differential privacy. The scheme integrates differential privacy with the peak-valley TOU model, optimizing pricing strategies while safeguarding user data privacy. Differentiated pricing schemes are implemented to influence users’ behavior of electricity consumption, encouraging off-peak consumption habits and achieving a balanced power load across the system. Experimental analysis assesses the data utility following the introduction of differential privacy, evaluating the operation effectiveness of the proposed mechanism. Results show that this scheme successfully achieves load balancing throughout the whole power network while protecting user data privacy.