This paper focuses on the demand for intelligent assistance in English business writing scenarios and proposes an intelligent assistance system for English business writing based on image recognition algorithm and language model. The system is able to quickly extract image information related to the writing topic through the similarity vocabulary matching technology combined with the image retrieval recognition function based on CBLSTM-Attention model. The language model is utilized to make accurate vocabulary recommendation and expression for the writing scene and user input content, and finally construct the overall framework of the intelligent assistive system based on English business writing. The system performs well in terms of vocabulary matching accuracy and writing efficiency improvement, with an average matching accuracy of over 90%. Students’ quality of writing is essentially improved with the help of the system in this paper. The actual case study shows that studying under the intelligent assistance system, the post-test scores of English business composition of the students in the experimental class increased significantly by 9.9393 points (P < 0.05) compared with the average scores of the control class, and it is obvious that applying the model of this paper to the classroom teaching can lead to a significant improvement in the performance of the students, which demonstrates the good prospect of its application.