with the development of computer technology, computer-assisted teaching has become an important aspect in modern education application. More and more educational software plays a major role on helping people learn the language. In this paper, we combine natural language process technology to multimedia technology, and develop a computer-assisted teaching system based on natural language process technology to further intellectualize. Firstly, we investigate on the reliability and feasibility of the application of automatic Chinese text segmentation technologies. And then we probe into the readability of Chinese text, and analyze the factors influencing readability from the perspective of vocabulary. Finally, we propose the readability determination technique based on statistics of graded vocabulary, and establish the standards for readability determination. The feasibility of this technique has been verified by experiments. The exploration provides theoretic analysis and experimental evidence for the application of computer-assisted teaching.