There are more and more web services used in collaborative learning, hence it has become increasingly important to locate proper web services in an accurate and efficient way. In this paper, an annexed algorithm, called eLSDAUS, is proposed to improve the existing semantic-based matchmaking algorithm. In the algorithm, we introduce and incorporate in a new factor -- User Satisfaction, which is the user's level of satisfaction about the result of service discovery. This algorithm allows users to take part in the process of e-Learning service discovery, and evaluate the result of service discovery. User's evaluation in the form of User Satisfaction is fed back to the system. Adopting an amendatory function that takes the User Satisfaction as an input, the system modifies the weights of each property of the advertise service, and then total match degree of service discovery up to best. Experiment indicates that the precision of the service discovery improved 5 percent as the number of advertisement services up to 1000. After learning for one week, over 93% students are satisfied with the e-Learning service discovery result.