The inrush current is the main factor that leads to misoperation for transformer differential current protection. This paper proposes a novel algorithm based on waveform cross correlation principle for distinguishing between magnetizing inrush current and internal fault currents. Either in the case of no-load switching or internal faults, a short sampling data-window is used to produce a normal sine wave as reference, and then calculate the normalized cross-correlation coefficient between the practical sampled waveform and the reference. In the condition of internal faults, the normalized cross-correlation coefficient is close to 1. But in the situation of inrush current, it is deviated from 1 and waves notably. Therefore, the calculated normalized cross-correlation coefficient is used to detect internal faults and inrush current. PACAD simulation results indicate that the proposed algorithm can identify inrush current and internal faults correctly and quickly, and it is helpful to supply theoretical basis for improving the performance of differential protection and optimizing protection scheme in transformers.