Abstract—Discrimination of bacteria is necessary because bacteria are remarkably crucial as they can contaminate in many possible ways such as food, water, air or environment, and lead to pathogenesis. In recent times, the technique of discrimination and counting bacteria are not only have a large expenditure of time and funds, but the discrimination method also required an expert in field. According to above, this research introduces an algorithm for solve those problems by using Staphylococcus aureus and Salmonella enterica as a pilot study. The stab technique, to inoculated bacteria in Luria-Bertani agar (LB agar) plates, was used for providing a good detection. In addition, the image segmentation technique was invited to train the separate colony of each bacterium for machine learning algorithm. It took approximately 800 images of bacteria for training process. The algorithm had resulted in 98.75 % of accuracy in bacteria’s colony counted and 98.12% in discriminated between Staphylococcus aureus and Salmonella enterica from stab technique. The algorithm had resulted in 33.61 % of accuracy in bacteria’s colony counted and 41.73% in discriminated between Staphylococcus aureus and Salmonella enterica from spread technique