The invention discloses a sampling discrimination method for sensor data exception, and belongs to the technical field of detection. According to the method, the fault of signal disappear and the faults of keeping constant or exceeding the upper and lower limits, that are not judged by the rules and that the measured targets cannot be measured can be discovered from the sensor signals, so that thereal monitoring data is obtained. According to the method, the control range is calculated from the global perspective, a data sample cycle is long, and the influence of instability of new data is smaller, so that the obtained result is more stable and accurate. According to the method, the mature technology and tools of an SPC method are utilized, and dependence on other reference standards is not needed, so that the sensor fault can be timely and effectively judged and alarmed, the equipment shutdown and sensor maintenance or replacement are carried out, the requirements for control, safetyand the like of the measured target are met, and the losses, such as production out-of-control, equipment faults, etc., caused by the faults, are avoided.