Abstract: This paper presents an intelligent method of eigenvalue extraction for the systems with a great deal of mixed and stochastic disturbances. Through searching features of waveform signal of objects, this method can identify and extract feature signals which can indicate real state of the measured objects and restrain the influence of disturbance in field signal. As a result, the testing and monitoring accuracy is greatly improved. Theoretic description and application of the expression of object features, and formation and extraction of feature signal are presented in this paper. Successful application of this method to practical projects shows that it has a good ability to process disturbance signals. Especially, it is very suitable for the cases of system analysis and monitoring under the influence of field mixed disturbance.