To moderate the behavioral variability of mouse dynamics, the dimensionality reduction based approach was proposed. Mouse dynamics is the process of identifying the user based on their mouse operating behavior i.e.) how the user may operate the mouse on a particular period. The mouse dynamics data set includes mouse operation, co-ordinates axes and time stamp value, from the collected dataset, the schematic and motor-skill features were extracted to obtain feature vector. Then the dimensionality reduction based approaches were applied on feature vector space. Authentication task is done by SVM (Support Vector Machine) to identify whether the input sample was legitimate user (or) imposter. The test result proves that the proposed method Isomap (Isometric feature mapping) with SVM provides better performance than existing system KPCA.