Adaboost as a kind of iterative algorithm, is a kind of important integrated learning technology, can be randomly surmised with slightly higher prediction accuracy than usingweak learning enhancement study for high prediction precision of apparatus, it’s is very difficult in strongly constructing learning, it provides an effective new way of thinking and new methods for the design of the learning algorithm. As a kind of algorithm framework, Boosting can be applied to almost all the popular machine learning algorithm to further improve forecasting precision of the original algorithm, and its application is very extensive and has a great influence. Based on Adaboost, 90 students’ oral English at a university as weak classifier samples, via the iterative calculation, get a high reliability of strong classifier, and it’s concluded that the prediction result with student can evaluate the prediction accuracy of the algorithm for results compared with the actual results.