This paper introduced the theoretical foundations of the iterated func tion systems (IFS), and then all the parameters of IFS were computed. An adaptiv e fractal algorithm for the vibration signal data compression and reconstruction was presented based on the IFS theory and the fractal property of vibration sig nals of rotating machinery. The characteristic of the adaptive fractal algorithm is that the length of the subsection is decided by the error threshold accordin g to the property of vibration signals. Finally, the compression ratio of the vi bration signal from an engineering application and signaltonoise ratio of th e reconstructed signal were calculated. The results show that the proposed algorit hm is practicable.