The paper represents an attempt to apply the general principles of modelling vegetation dynamics under climate changes to a study of the long-term vegetation dynamics in the forest-steppe zone of the European territory of Russia, with a purpose to forecast under special climatic scenarios. An original technique is used to construct a Markov chain as a model of vegetation succession. The technique emanated from gebotanic knowledge generalized as a scheme of successional transitions with estimates of the average duration for certain stages of succession. Whenever the knowledge related the stage duration to certain (climate-sensitive) factors of the environment, the fundamental potentiality arises to model the temporal course of succession as a function of a given scenario for how the key factors change. In the formal terms, the model represents a random chain of the Markov kind with a finite number of states and discrete time of transitions by the given scheme. Relative square distributions of succession stages under concern at any time moment (within an adopted scenario) appear as the model outcome (forecasts), as well as estimates of the attainment time for certain states of the vegetation in the territory under study. A method is proposed to describe dynamics of the phytomass production and stores (and the corresponding model trajectories are obtained) for a given scenario.