In order to further develop the strategy and methodology of Dynamical Analogue Prediction (DAP), the physical basis problem associated with the DAP is primarily investigated in the present study based on previous nonlinear dynamical achievements. Firstly, the relationship of “analogue prediction-low-frequency flow regimes―predictability” is fully discussed. Analyses show that such a relationship is not only suitable for statistical analogue prediction, but also for dynamical analogue prediction. Furthermore, by contrasting with initial ensemble forecast and conducting case analyses of dynamical predictions on different time scales in real atmosphere, it is primarily documented that there exists evident analogy between dynamical prediction results and their situations of prediction errors when initial values are analogical to each other. Moreover, three classes of analogy problems are further summarized and the physical basis of dynamical analogue prediction is induced: Under the circumstance that the analogy at initial states is durative, not only dynamical prediction results but also their situations of prediction errors will be analogical to each other.