This paper proposes a new method of deterioration trend control based on information theory for condition monitoring of plant machinery. The Information Divergence(ID)and Kullback-Leibler Information(KI)are introduced in time domain and frequency domain for the purpose. By the theoretical analysis, it is shown that the failure detecting sensitivity and the failure distinguishing ability of ID and KI are higher than conventional ones in time domain. Based on information theory, the criteria of ID and KI for condition judgement are induced and the simulation results are shown to verify the efficiency of this new method.