A novel double gray image encryption method is proposed based on compressive sensing (CS) and interference principle. The 50% image data of two gray images are random extracted by the proposed method, these data are combined to form a synthetic image (SI), and then it is hidden into three phase only masks (POMs). One of the phase masks is produced by random function and the other two are obtained by analytic method. For decryption, the diffraction fields of the three POMs are superposed by employing the beam splitters. The intensity of the complex field, namely the SI, is captured by CCD camera, and then the information of two original images is extracted from the SI. Although only 50% fragmentary data are extracted, the subsequent CS reconstruction will retrieve a high quality image from the fragmentary information. Compared with the earlier interference- based method, the proposed approach is a process of employing resolution algorithm completely and time-saving since no iterative algorithm is involved in the encryption process. Moreover, the silhouette problem existing in the earlier method is resolved by the proposed method with higher security. Simulation results are presented to support the validity of the proposed approach.