In order to reduce the sensing data transfer frequency and node energy consumption in real-time cold-chain logistics monitoring system for agri-food based on wireless sensor networks, a sensing data compression method based on statistics process control was presented. The D-X-RS algorithm that fits for the sensor nodes’ data processing capability was designed. The data compression performance of D-X-RS algorithm, threshold algorithm and K-means algorithm by cold-chain logistics environmental simulation in laboratory was compared. The result showed that the compression cost per point (CCP) of D-X-RS algorithm was close to the best algorithm both in temperature fluctuation and steady states. The algorithm had good balance and flexibility. 