The widespread use of distribution-level phasor measurement units (D-PMUs) in digital distribution networks has introduced significant pressure on communication and storage systems due to the high precision and data transmission rates of these devices. This paper presents an adaptive compression algorithm for D-PMU data from edge computing devices in digital distribution networks. Utilizing a filter-based swing door trending compression algorithm combined with state discrimination methods, the proposed method applies distinct compression strategies for stable and fault operational states, achieving adaptive data compression. Validation of the algorithm’s effectiveness was performed through testing on system data during both stable and fault conditions. The results demonstrate that the proposed method effectively reduces the load on communication and storage systems while maintaining data accuracy and real-time performance, thereby proving its feasibility and efficacy in practical applications. This approach ensures that critical data is preserved during fault conditions while efficiently compressing data during stable operation, thus enhancing the overall efficiency of digital distribution networks.