Marine hydrokinetic (MHK) turbines extract renewable energy from harsh marine environments, where biofouling and corrosion acting on turbine blades will affect system performance and lead to progressively increasing damages. Thus, accurately estimating a blade's remaining useful life (RUL) is important to achieving condition-based maintenance to ensure secure and reliable operations of MHK turbines, and the reduced cost of hydrokinetic power. In this paper, we propose a new RUL estimation method based on adaptive neuro-fuzzy inference system (ANFIS) and particle filtering (PF) approaches, establishing a relationship between blade imbalance faults and the produced power signal. The ANFIS is trained via historical failure data, and it constitutes with a m-order hidden Markov model to describe the fault propagation process. The high-order particle filter uses this Markov model to predict RUL in the form of a probability density function through collected normalized time series data. Results demonstrate the strong potential of the proposed approach for MHK turbine lifetime prediction.