Abstarct: Image analysis is an Image processing technique that aims to extract features or information from images. Image analysis in medicine has a special place because is a basis for disease di-agosis for physicians. In addition to diagnostic imaging technique that different parts of the body are covered, sometimes it is necessary for the diagnosis is used of the patient's blood. In order to separate blood components using a machine called Electrophoresis. Electrophoresis is mobility of the particles in a liquid under the influence of an electric field uniform. Serum protein electrophoresis and hemoglobin electrophoresis test are the most important laboratory tests that used now widely in the medical diagnostic laboratories in order to diagnose certain diseases of the blood. Images obtained from electrophoresis are analyzed by hematologists and electrophoresis im-age analysis software. The important part of the analysis is separation of the bands that this can be done manually. Also, lane curve position moved by a hematologist. Thus electropho-resis software can’t automatically identify band’s separators. In this case there is the risk of mistakes due to human error. In this dissertation introduced technique that identifies automatically location of the bands on electrophoresis images that includes pre-processing, lane extraction and lane analysis. Prepro-cessing stage to calculate the rate of change in the vertical profile picture electrophoresis of lane for identify and extraction area of the lanes. In lane extraction stage, the determination of appropriate threshold is done using Otsu method then extracted all of lanes. In lane analysis stage, histogram stretching method used to improve the contrast of the lane then the position of the separators can be determined. Then consider the relative distance between the bands, optimized position of the separators is calculated. After specify bands locations, values of the bands are determine. Proposed algorithm done on images obtained from Labs medical diagnostics Shahrekord, Is-fahan and shahrood cites. By comparing the results of the algorithm and the results provided by the laboratory diagnostic rate of 75.6% is obtained for proposed algorithm.