Colon cancer is the second top reason of cancer-causing deaths worldwide. It is very lethal, and so early detection can increase the chance of survival of patients. An automated system using deep and machine learning technology are used to identify colon cancer in less time with more accuracy and at lower cost. We present a thorough review of current research on colon cancer, analyzing models and algorithms in machine learning, deep learning, and transfer learning. We found best algorithm by comparing the prediction results to detect colon cancer and found that MobileNet transfer learning model and CNN with MobileNetV2 model gives better accuracy and also the more clear and sharp image leads to better accuracy. This technology can identify cancer at early stages, and it will head for better treatment of the colon cancer and so it lowers the mortality rate. Lastly, future research can be done on newly available larger datasets and different varieties of datasets and also on different algorithms to enhance accuracy.