In an alternative stable state (ASS) ecosystem, tipping points lead to abrupt and possibly irreversible shifts between alternative ecosystem states. However, there are only a few systematic studies on the ASS theory at present, and how to use a variety of early warning signals to predict the regime shift in the ecosystem at different scales is still a difficult problem. Besides, critical transitions are difficult to predict because the state of the system usually shows little change before the transition. In this review, firstly, a more integrated definition of ASS has been put forward through reorganizing of ASS theory from different previous literatures. Alternative stable states are defined:under the same external force or internal force driving, the structure, material composition and energy flow of the ecosystem will change into two or more different stable states, which are maintained by negative feedback loops. Then, the relevant concepts of ASS theory were systematically summarized by the model of the grazing animals on pasture land. Furthermore, ASS theory has been originally applied to interpret succession and panarchy theory in the ecosystem. Secondly, the principles, advantages, disadvantages, as well as application conditions of the early warning signal of regime shift in an ecosystem were analyzed, and the research methods of ASS theory at different ecological scales were summarized. The mathematical reason for critical slowing down leads to an increase in the autocorrelation and variance of the fluctuations in a stochastically forced system approaching a tipping point is illustrated by the autoregression model. Finally, the current research problems in the field of ASS and the focus of future research were put forward. The review results show that:(1) combining different early warning signals could provide a more accurate prediction of the tipping point by quantifying the probability of correct early warning signals and weighting the proportion of wrong early warning signals. (2) The quantitative experimental observation experiment is suitable for small-scale research, while larger-scale research usually uses a simplified model to simulate the results. Moreover, choosing the correct scale is very likely to improve the reliability of the early warning signal in predicting the regime shift. (3) The application of ASS theory would be important for ecosystem restoration by studying the critical transition and feedback control mechanism combining the trait-based characteristic indicators and evolutionary dynamics. (4) Applying ASS theory into the practice of ecological protection and management policy is very promising. This study provides scientific support for the further research of ASS theory and practice.