The abnormal monitoring of sensor data in stadiums is of great significance to the deployment and design of IoT technology. Because the nodes that need to be monitored in the stadium are scattered, the monitoring sequence of traditional sensor nodes is chaotic, resulting in large deviations in detection results, and when the number of nodes is large, the remaining energy is low. In order to establish an effective and intelligent stadium temperature sensing device, This paper proposes a method for monitoring anomalies in stadium perception data based on edge computing and time series. According to the stadium sensor network structure based on edge computing, it is divided into cloud layer, edge layer and bottom layer, with the goal of minimizing energy consumption. Analyze mobile edge computing elements and design sensor data storage and distribution solutions. The time series is used to mine the anomalous data of the stadium. On this basis, taking the temperature of the stadium as an example, the abnormal monitoring result of the sensor data of the stadium is obtained. In order to verify the effectiveness of the proposed method, a simulation experiment is designed. Experimental results show that the detection accuracy of this method is high, and the residual energy of nodes has no obvious downward trend. The experimental results verify the ideal application performance of the proposed method.