In the past, most automatic evaluation methods for English writing were based on statistical characteristics such as grammar, spelling, and vocabulary. At present, they mainly rely on manual annotation of corpus and the development of scoring rules, resulting in high evaluation time costs. This chapter aims to provide feasible solutions for English writing education and evaluation, in order to effectively improve the efficiency of cultivating and evaluating students’ English writing abilities. This chapter studied an automatic English composition evaluation method based on TF-IDF (term frequency-inverse document frequency) and text similarity algorithm. By calculating the similarity between student compositions and standard compositions, the composition level of students was evaluated. This study verified the effectiveness of an automatic evaluation method for English writing based on TF-IDF and text similarity algorithm through experiments. The experimental results showed that the average time interval for evaluating the writing style, vocabulary, tone, originality, and structure of 355 English compositions in control group 1 and control group 2 was between 1.5 and 2 min, and the average evaluation time of the experimental group was within 1.5 min. This method can greatly reduce the evaluation of student English writing quality by English teachers and improve the efficiency of the evaluation. The application of TF-IDF and text similarity algorithm for automatic evaluation of writing not only facilitated the teaching of English writing by college English teachers but also had a significant impact on the English writing level of students.