While the existing de-noising algorithm requires prior knowledge of vibration signals, a new adaptive de-noising algorithm is proposed based on sparse coding and dictionary learning (DLSDF). Depending on the essential attribute of different signals, the optimal dictionary of data-driving is learned from the raw data. The orthogonal matching pursuit algorithmworks out the sparsest coefficients. Then, the de-noised signal is reconstructed using sparse coding and the optimal dictionary. Simulation and experimental results show that the algorithm based on sparse coding and dictionary learning is adaptive, and de-noising is stronger than the existing one.