In response to the problem of a large number of test flight fault problems and the inability to efficiently utilize their information value, a recommendation method for similar fault problems based on the index system of flight stages, ATA, fault text, CAS and OMS information was proposed. Similarity calculation methods were proposed for each indicator based on their characteristics. In addition, a method is proposed to automatically extract CAS and OMS information from fault text descriptions using regular expressions. Finally, an example was used to validate the recommendation algorithm that only selects flight stage, ATA, and faults text as indicators, as well as the algorithm that only uses TF-IDF method. The algorithm proposed in this article has improved 25% and 28% in accuracy, 27% and 31% in recall. It can provide reference for the recommendation of similar flight test fault problems.