In the context of the energy internet，traditional active power balance regulation products fail to meet the reliability，flexibility，and low-carbon demands of new power systems.To address the problems that domestic regulation products are single in configuration and ill-equipped to handle the increasing volatility of generation，the concept of balance demand is introduced，and an active power balance service product system based on signal time-frequency domain analysis is proposed.The method uses K-shape clustering to identify typical daily balance demands，applies ensemble empirical mode decomposition （EEMD） for multi-level decomposition of regulation signals，and standardizes product design through time-frequency domain analysis.Quantitative analysis is conducted using frequency regulation and reserve products as examples to demonstrate the effectiveness of the proposed method.The results show that this product system design approach facilitates optimal allocation of regulation resources and reduces market costs.Guidance is provided for the hierarchical classification of regulation capacity in regional power systems，the determination of product numbers and response time requirements，and the design of active power balance service product systems tailored to regional conditions.