This paper proposes a new mobile phone user authentic scheme. It discerns a particular user by the smartphone sensor data which is collected when he picks his phone up from table. The data from phone sensors which collected by preprocessing method, it includes denoising, interception and normalization. Precondition the sensor data with denoising and signal truncating technology, then normalize it, finally classify the data by Dynamic Time Warping (DTW) algorithm. This scheme also has the intelligent anti-theft function. Experiments showed that this scheme achieved good results: False Acceptance Rate (FAR) is 13.1%, False Reject Rate (FRR) is 11%.