When the equipment condition is being deteriorated, some transients of occur. Thus processing and transforming of signal will be a good tool for the development of novel condition monitoring scheme for equipment. The change of spectra is caused by abnormal and failure of equipment. Extrating of the spetral features becomes eflicient tool of condition monitoring and failure diagnosis of equipment. With condition monitoring and failure diagnosis, maintenance is scheduled only when it is required, but before failure occure. In this way, downtime is minimized, maintenance cost are reduced and life of equipments is extended. This paper describes basic principle and modeling method of extraiting of frequency of peak value by non paramter and paramter. Its application on diagnosis of equipment is introduced.