<p>The development of natural language processing technology can help solve language barriers between humans and computers, improve the efficiency and convenience of human–computer interaction. This article utilizes natural language processing and decision tree algorithms to classify and correct error types in computer English writing, in order to improve learner’s writing proficiency. This article uses natural language processing technology for data preprocessing to analyze and understand text more accurately. Subsequently, decision tree algorithms are used to train the samples and construct a classification model. By analyzing and judging the features of the sample data, a series of decision rules are generated. By continuously adjusting the parameters and structure of the decision tree, the model can classify samples more accurately. Finally, the constructed classification model is used to classify and correct new computer English writing. The model will identify the types of errors based on learner’s writing characteristics and existing decision-making rules, and provide corresponding correction suggestions. The experimental results show that the method proposed in this paper performs well in classifying and correcting error types in computer English writing. By correcting learner’s writing, it helps them avoid common grammar and vocabulary errors, thereby improving the quality of their writing.</p>