Abstract: [Purpose/significance] In recent years, the loss of users has become an important factor hindering the development of knowledge sharing community. Evaluating the contribution of community users from the perspective of user engagement will help attract potential users, increase user stickiness, and reduce user loss. Accordingly, this article evaluates the contribution of community expert users from the perspective of improving user engagement. [Method/process] First, we built an expert contribution evaluation system based on customer engagement theory, then used Baidu Know Daily as the research scenario to obtain expert contribution behavior data under its selected columns, and then used the entropy weight method to calculate the entropy weight of each evaluation index. Finally, the expert contribution degree was obtained according to the TOPSIS evaluation method. [Result/conclusion] The resuls show that,different types of engagement behavior have an impact on expert contribution. Also, evaluation results overlap with the popularity of platform users nearly in half. Meanwhile, the distribution of expert contribution is unbalanced, showing the phenomenon of“Matthew Efect”". In addition, experts with different user characteristics (such as expert types, etc.)also have different contribution level. Finally, this article proposed a number of strategies based on the perspective of user engagement to help the knowledge sharing community improve the contribution level of expert users.