This paper attempts to do part of the exploration work on the application of big data in the study abroad industry with semantic technology, and builds a potential study abroad customer data collection system based on semantic technology, and mines the potential study abroad customer data from the corpus big data of public social media. This paper presents a disambiguation algorithm based on "HowNet". The algorithm uses the four relationships contained in the semantic origin of words to discuss the timeliness of carrying out online ideological and political education for college students through new network media and various forms of communication. It shows that the education platform has indeed achieved a positive interaction with students, allowing students to be educated and inspired, thereby improving the pertinence of college students' ideological and political education. Calculate the correlation between words and the correlation between words and their contexts, and then achieve the purpose of word sense disambiguation.