Progress in Machines and Systems Yu Yong-Wu 6 2 2017 http://www.dline.info/pms/fulltext/v6n2/pmsv6n2_3.pdf With the rapid development of computer, people are increasingly demanding the quality of the mainstream cloud platform which is main forms of computer industry. In cloud monitoring system, picking up information, monitoring data anomalies, abnormal transmission and alarm should be vigilant. By studying the theory of cloud platform monitoring technology, this paper uses FASTMASOD algorithm to design the node machine, and uses the host computer to test the anomaly detection system of cloud platform in the context of multi-attribute information. Finally, the ROC curve is used to verify the effect of the algorithm in the anomaly detection system. The research of this paper not only lays a theoretical foundation for the anomaly detection of multi-attribute information, but also has very important significance to the research of cloud platform.