With the widespread proliferation and application of high data sampling rate device-Phasor Measurement Units (PMUs) in modern smart grid, large volumes of data are produced in the wide-area monitoring system, which results in increasing challenge in real-time analysis, data storage and data transmission. This paper proposes an improved lossless compression technique which is fully based on the analysis of PMU data structure and consists by three stages algorithm for handling massive data. In the first stage, adjacent coordinate data of one period are operated subtraction and the waveform difference method is adopted to simplify data of other periods. Then, adaptive dynamic Huffman algorithm is selected to compress the processed data. Finally, Run Length Encoding method is employed to decrease furthermore the redundancy. The performance of the improved approach was evaluated under simulation using field PMU data. Experimental results under both dynamic and steady states presented at the end of the paper prove the efficiency of the proposed technique.