With the in-depth development of economic globalization, the connection and communication between China and other countries in the world are increasingly deepened, so the task of cultivating translation talents in various languages is particularly important. In the teaching process, teachers mainly focus on explaining theoretical knowledge and lack of classroom interaction, resulting in a boring and dull classroom atmosphere. At the same time, it does not pay attention to the practice of translation after class and often pays attention to interpretation and neglects written translation, which leads many students to think that Japanese translation is difficult and boring, which greatly reduces students’ enthusiasm for learning. In view of this situation, the application of multimedia technology in Japanese translation teaching proposes that the application of multimedia technology in Japanese teaching can improve the interest of students in Japanese learning and improve the teaching effect of Japanese teaching. Firstly, this paper constructs a logistic semantic ontology model based on automatic calibration of Japanese translation, extracts Japanese semantic features through this model, and performs relational mapping on the extracted Japanese semantic features. Starting from each aspect, optimize the automatic calibration algorithm of Japanese translation, find the best semantic relevance feature in each sentence, and realize automatic optimization, thereby improving the automatic calibration of Japanese translation machines and the registration of subject words.