Along with the development of technology, space weather activity becomes a very important thing in science. It’s caused by whether activity that occurs in space can affect the activities of life on the earth. Therefore, it’s important to be able to detect whether events in space, including solar flares. We believe that there has not been a single solar flare prediction study that did a prediction using PROBA-2 SWAP data, because flares are difficult to catch at that frequency. Nearly all previous researches have been focused on SOHO / MDI and SDO / HMI satellite. If the two satellites can’t capture the image for some reason, then PROBA2 SWAP satellite imagery can be an alternative. This research is aimed to implement image processing and machine learning methods on SWAP PROBA2 satellite imagery to predict event numbers of solar flare. The machine learning algorithm used is random forest, while the segmentation algorithm used is seeded region growing. This solar flare prediction research using SWAP PROBA2 satellite imagery produced the best f-measure value of 0.897