Abstarct: Nowadays , one of the major problems in developed countries is the issue of infertility and its treatment . The problem that left many couples involved . In examining the infertility problem, first strategy being investigated is study of the problem in the male side couple . According to the different causes that may be involved in male infertility ,test and analyze his sperm is the first step to study of specialist doctors. Sperm analysis in the male couple is examined in several aspects , including the study of the form of a sperm , as well as the analysis and evaluation of the strength of the motility of sperm. Of the two mentioned factors, the second is the most important factor. Since the second one has most importance if it will necessary to vitro fertilization . due to the structural nature and motion characteristics of sperms , the use of accurate and powerful tools for analysis and evaluation of sperms motility has particular importance . Nowadays , artificial intelligence and especially machine vision are used as one of the most important tools in different fields of medicine . In this research , we use microscopic video frxames of sperms. First we using kalman filter and Hungarian algorithm to trace them and extract the original and conventional kinetic parameters. In the second step of this study , using existing algorithms in machine learning domain. In order to analysis , evaluation and classification of fertility performance of each sperm in terms of their motility power . In our activity , we used a support vector machine to classify the sperms in terms of motility power in two existing classes . According to the results , the final output from classification using support vector machine method , with accuracy of 88.9 percent , could predict the class related to sperms . The importance of this research is to carry out the classification operation for all the main features available for the sperms . therefore, it is shown that , for assessing the quality of motion , parametric methods can be introduced to classify them using the parameters from motion detection .