With the continuous improvement requirements of reliability and security for mechanical equipment,accurate extraction of equipment fault development trend of degradation characteristic information and establishment of effective fault prediction model is the key problem to improve reliability of equipment operation. HMM has double random process properties that could describe the hidden state and observed state,it is similar to the equipment degradation processes that become a hot research topic now. On various methods based on the fault identification and diagnosis technology, the equipment fault prognostics methods based on hidden Markov model and hidden semi-Markov model are summarized, advantages and disadvantages of each method are analyst and compared. Finally,its development trend on equipment degradation assessment and prognostics is pointed out.