There are many cases in which the common vibration signature analysis methods are unable to diagnose the phenomena resulting from the initial damages in a mechanical system. Since the initiation of a damage in complex system like gearboxes and rolling element bearings, appears in the vibration signal generally as a low-energy and short-duration phenomena, the traditional methods fail to detect these small changes in the vibration signal. The methods based on the wavelet analysis which have been introduced recently, are capable of analyzing a signal both in the time and frequency domains simultaneously. Among the various wavelet systems, the biorthogonal wavelet systems have unique properties which provide the flexibility of the analysis method, However, up to now the biorthogonal wavelet systems have not been used in the area of the vibration signature analysis. In this paper, for the first time, the application of biorthogonal wavelet systems are developed to damage detection of mechanical systems through detecting the transient and the other hidden phenomena in the vibration signals. For this end, some biorthogonal wavelet and scaling functions have been designed and the developed biorthogonal wavelet systems have been used to analyze some simulated vibration signals. One of the most important properties of the mentioned systems is their capability to analyze and to reconstruct the complicated signals which contain harmonic and transient components, faster and more accurately. In order to show the advantage of biorthogonal wavelet systems over the orthonormal wavelet systems, some simulated vibration signals have been analyzed using the mentioned techniques. The result show the great capability pf the biorthogonal wavelet systems to diagnose the transient and low-energy phenomena in the vibration signal as to remove the noise from the vibration signal.