Abstract: This paper mainly studies the pricing problem of "photographing to make money". Based on the K-means cluster analysis, the pricing model of the task pricing and the number of members around the task, the credibility of the members, the time when the members start the scheduled tasks and the location of the task distance are established. The logistic regression function is used to predict whether the task can be completed to evaluate the completion of the model. Finally, the task pricing obtained by the algorithm is compared with the pricing given by APP, and the correctness of the model is verified. In addition, by establishing a packaged task points is increased from 70.5% before packaging to 80.3% when the total task price is unchanged. The packaging has a significant improvement on the task competition.