Abstract: Key index analysis state was described qualitatively by using cloud theory and multi-sensor weighted mean method in this paper.Firstly, the data got from the sensors in natural state were tested conformably by using distributing graph method.Secondly, we can use the multi-sensor weighted mean method and get the smallest data confluent value of the average error in natural state.And then compared the key index data in any time with foregoing confluent value and the warp value was gained.At last we can construct five remark sets based on cloud theory, and combining the warp value we can describe the key index analysis state qualitatively.The experiment result indicates the validity of our method.