In order to solve the malicious code of cloud computing architecture in various forms of intrusion damage, can not be found in time and maintenance caused by cloud computing architecture safety performance is reduced, not the normal use of the problem, based on a BP neural network intrusion detection system, realize the calculation of automatic monitoring in the framework of the invasion of the malicious code on the cloud, have an important role in the direct and timely monitoring of security computing architecture intrusion malicious code and effectively increase the cloud system; using STM32F103ZET6 as main control chip of the main control unit of MUC, and the EZ-USB FX2 USB2.0 control chip is connected with each module; using LM2575 series voltage regulator, power supply system; software design process, the output of the calculation of the invasion of the malicious code the value of using BP neural network method, reduce the monitoring error; the experiments show that the system can achieve the cloud The automatic monitoring function of intrusion malicious code in the architecture has the characteristics of strong expansibility and easy operation, and has important application value to the security of the cloud computing architecture.