Journal of Information & Systems Management Feng Qi, Zhifang Liu 8 2 2018 https://doi.org/10.6025/jism/2018/8/2/55-62 http://www.dline.info/jism/fulltext/v8n2/jismv8n2_2.pdf In E-business, the study for retrieval method for fuzzy customer requirements is very valuable. By modeling the fuzzy customer requirements, the retrieve model is more adapt to the E-business engineering fact. Refer to the special customer groups, the customer requirements can be translated into the certification and judgment for the attribute parameters of digital product feature. According to previous experience or the condition of using the digital product, customers can directly put forward quantitative or qualitative request to the attribute parameters of existing digital product feature of digital product series. In allusion to insufficiency of digital product retrieval based on similarity, the Neural Network is introduced into design of case retrieval and a model based Neural Network is proposed. By studying modeling NN and fuzzy requirements information, the mapping network from the parameters of digital product feature to the candidate of digital product has been set up. Rapid response to fuzzy individuation requirements of customer has been realized. This method, as the key technique of Ebusiness, is applied to gearbox design of enterprise, which can greatly shortened design cycle of no-standard digital product in digital product series.