The aim of this research is to develop an intelligent service system to receive, manage, and serve text messages sent by system users asking for files stored in a system database. The proposed system needs to understand the received text message and takes the decision about the required files. The system handles files and a corresponding computer program and a corresponding computer-readable storage, which can be used for handling e-mail attachments and sent the requested files to the user email. To improve the designed text message recognition system, the combination of two methods have been used, a supervised learning multilayer neural networks trained with back propagation and Soundex algorithms. Experimental results show that the combination of two classification techniques to recognize and manipulate the received text message improves the classification accuracy.