Effective fisheries stock assessment and management is the key to achieving sustainable exploitation of fisheries resources. Traditional fisheries stock assessment and management is based on the assumption of stationarity, which considers key processes of population dynamics as time invariant. With increasing impact of climate change and human activities, these key processes may change over time, leading to non-stationary population dynamics and challenging the traditional assumption of stationarity. Based on the recently published research on this topic, we reviewed the mechanisms of non-stationary population dynamics and their impacts on fisheries stock assessment models, biological reference points and fisheries management. The results indicate that in the light of non-stationary population dynamics, both stock assessment models and biological reference points need to consider the temporal variation of key population processes, either by including environmental covariates or by time-varying model parameters, to deal with the impacts of non-stationary population dynamics on fisheries stock assessment and management. We suggest that resilience-based fisheries management is an important tool to cope with non-stationary population dynamics and maintain a stable and healthy fisheries social-ecological system.