An essential challenge in the field of education is predicting student performance since it enables teachers to spot pupils who could struggle academically. Machine learning makes extensive use of hybrid optimization methods to boost the precision of prediction models. Large datasets may be processed in parallel using the Map Reduce programming style, which makes it the perfect platform for big data performance prediction applications. The prediction problem is split into two parts in a map-reduce architecture using a hybrid optimization technique. The data is split up and spread among many cluster machines in the initial stage. Each division is subjected to a machine learning method via the map function, which produces a series of intermediate outcomes. The reduction function integrates the intermediate outcomes in the second phase to create the final forecast.