Abstract: According to the running characteristics of network crowdsourcing contests and related research of crowdsourcing performance, this paper applied the multiple linear regression method, and studied the factors that affect the performance of public participation from competitors' participation numbers, tenures, participation interval and other behavior factors. Through linear regression method, it identified the key variables in affecting individual performance of network crowdsourcing contests. The results show that competitor' tenures and submitted entries have significant effect on their performance positively; competitors' rankings and participation interval have significant effect on their performance negatively. Moreover, competitors' participation numbers moderate the relationship between tenures and performance negatively, and the relationship between participation interval and performance positively. In addition, competitors' participation interval moderate the relationship between tenures and performance negatively too.