Motion artifact distorted ECG signal is a commonly known issue in mobile ECG recording. In this paper, we present a better method to replace existing filtering technique while preserving distortion free ECG signal. A belt-type ECG measurement system to measure ECG signal in daily life has been developed. A 3-axis accelerometer sensor is installed on the ECG measurement system to obtain activity data. Activity status is classified using fuzzy rule base classifier. Then an optimal adaptive filter using empirical mode decomposition, EMD is proposed for motion artifact removal. EMD is used to decompose motion artifact from the ECG signal and finally preserve a distortion free ECG signal at the end of the result.