The analysis of cancer cells at the level has been studied. However, it was difficult to clarify the characteristics of cancer cells such as the speed of becoming worse, the efficiency of the medicine on the cell level. Recently, Laser Scanning Cytometer(LSC) which measures the data of cancer cells has been developed. Dr.Furuya et al. tried to analyze the characteristics of cancer cells using to the amount and the cohesiveness of protein. However, they could not find the efficient method for lassification of cancer cells. In the paper, we try to classify cancer cells using the amount and the cohesiveness of protein by Self-Organizing Map (SOM). The data used in the classification is the image data created from the amount and the cohesiveness of protein of cancer cell extracted from the patient by LSC. The result shows the probability that SOM would be able to classify cancer cells by those protein information.