With the acceleration of globalization, international education has gradually become the core content of higher education. In this context, students have encountered lots of intercultural challenges, making the evaluation and improvement of their intercultural communication competence (ICC) a key issue. It is difficult for traditional evaluation methods, such as questionnaire surveys, interviews, and teacher evaluations, to meet the needs of current educational institutions because they are subjective, time-consuming, and inefficient. After describing the ICC evaluation process of students under the international education management model, this study proposed a new automatic evaluation method based on the dual-scale convolutional neural network (CNN) model and the bi-directional long-short-term memory (Bi-LSTM) model of characters and words. The new method evaluated students’ ICC more accurately and objectively, providing a new perspective and tool for future research and educational training.