In view of the problem that TCM diagnosis and treatment data are complex, diverse and lack unified standards, this paper introduced computer information technology, combined with NLP and LSTM models, to improve the intelligent analysis ability of TCM diagnosis and treatment data, and provide scientific support for clinicians through an auxiliary decision-making system, thereby improving the efficiency and accuracy of diagnosis and treatment. First, this paper collected and integrated a large amount of medical records, treatment records, prescription information and other data from different TCM diagnosis and treatment platforms and medical institutions, standardizes data in different formats, and used natural language processing (NLP) technology for semantic analysis and data cleaning. Then, it built a classification and prediction model based on LSTM (Long Short Time Memory) to realize intelligent diagnosis and treatment recommendation generation for common diseases in view of the complex symptoms and individual differences unique to TCM diagnosis and treatment. Finally, based on the previous analysis and model results, an auxiliary decision system was developed to provide doctors with auxiliary decision suggestions in the treatment of complex symptoms and medication regimen recommendations during the diagnosis and treatment process. The results showed that after the use of the intelligent auxiliary decision system, the diagnosis and treatment time reduction rate was between 21.8% and 23.0%, and the accuracy rate reached 88.0%. Compared with traditional manual analysis methods, the efficiency of TCM diagnosis and treatment data processing has been significantly improved after the application of computer information technology. The introduction of computer information technology provides a new solution for the intelligent analysis and decision-making support of TCM diagnosis and treatment data, effectively solving the problems of complex, heterogeneous and insufficient standardization of TCM data.