The method of Empirical Mode Decomposition(EMD) can filter and denoise the nonstationary microseismeic signal.But there exists serious end effect in its application which could affect the decomposition results. For improving the end effect, the wave extension of microseismic data endpoint based on LMS adaptive filter was proposed to the low frequency characteristics of microseismic signal.Trained by the internal data, LMS adaptive filter extend the wave from the endpoint to the left and right sides.The minimum and maximum that were in coincidence with internal signal trend were obtained in extension process and the extension length was determined adaptively by extension algorithm.The wave was decomposed by EMD method and mode components were acquired.For correcting the error brought by the interpolation algorithm, he residual signals were extended again.According to characteristics of the signal frequency, the mode components were reconstructed selectively in order to denoise.The adaptive LMS fiter extension method is verified by the simulation data and the microseismic monitoring data from coal mine, the results show that the method can obviously reduce the end effect of empirical mode decomposition, and have great effect especially to microseismic signals of low frequency.