This paper classifies the data of the original data set by random forest algorithm, selects the nodes in the attribute space for iteration, and gets the number of decision trees in the random forest. Based on the decision tree, the information gain rate in the sample set of student achievement is calculated, the spatial distance matrix of the sample set is defined, and the centroids of each cluster of the matrix indicators are separated to get the evaluation results of indicators in the student achievement evaluation reform as superior. The indicator weights of student learning achievement are evaluated through five assessment indexes, in which the teacher rating weight is the highest 10. It shows that active use of the Internet is conducive to cultivating and delivering excellent film and television production professionals to society.