The research process of optical functional glass materials involves long research and development cycles and low efficiency. Greatly hindered the development of optical glass materials. The emergence of machine learning technology has greatly promoted the development of glass materials science. By learning the laws contained in the data, learning and predicting new data from the huge and complex glass data has accelerated the research and development process of optical functional glass. This paper summarizes and demonstrates several types of machine learning algorithms involved in the prediction of optical glass and briefly introduces them. On this basis, it focuses on summarizing the important applications of these theoretical algorithms in glass research, including accelerating and improving traditional glass research methods, assisting glass composition-property correlation prediction, and suggestions for optical glass formulation design. Finally, the application prospects and future development trends of machine learning in optical functional glass research are analyzed and forecasted.