Footwear impression is a common phenomenon in crime investigation cas-es. It is frequently found where the crime is committed. To identify this impression manually is quite difficult and time consuming. Due to the var-ious qualities, inconsistency of the impressions and variety of footwear outsole designs, make it impossible to identify foot print impressions man-ually in certain cases. The purpose of this thesis was to implement a model for identification the impression from a known set of outsole prints. If the impression is not in data set, the implemented model can determine simi-larity between the crime scene impressions with the suspect’s shoeprint impressions. Forensic experts can investigate the sources of impression as well as the suspects which can help the prosecution. Different algorithms are investigated during the implementation of the model. An automatic image processing method to improve the quality of a query image as well as reference images has been studied. To evaluate the similarity between the known footwear impression and unknown one several features have been extracted from the images. The extracted features are limited to such geometrical shapes as lines and circles. The similarity between extracted features is evaluated using bin-by-bin distance and cross-bin distance using histograms named Features Distance Calculation (FDC). The implemented model compares the features with pre-computed features of database.