Abstract: The image analysis efficiency is one of the most important issues in digital speckle correlation method (DSCM) applied in geotechnical experiments. On the other hand, the displacement magnitude and direction of some types of geotechnical materials are characterized by slow changes. In order to improve the image efficiency of DSCM, an optimized method named “local direction search method（LDSM）” based on the progressive displacement characteristics of geotechnical materials is proposed, and its application prerequisites and key algorithm equations are also introduced. Functions of LDSM is implemented by programing in self-developed software PhotoInfor. The most common angle parameter scope from 30 to 45 degrees is given for the image search scope used in LDSM. The research shows that the image analysis results of LDSM are identical with those of the normal search method in DSCM. While the angle parameter in LDSM equals to 45 degrees, the analysis efficiency of LDSM is 1.3~1.6 times more than that of the normal search method. In other words, the analysis efficiency is improved by at least one time and is stable with the increasing analysis points on images. LDSM for DSCM has important practical values in geotechnical experiments.