This thesis deals with the methods used in analysis of strategies. Although the theory is generally given, the examples are limited to the analysis of business strategies. All of the described methods are based on the assumption that the observed process can be described by homogeneous Markov chain. Another important assumption is the assumption of the ergodicity of the observed Markov chain so that it would have sense to observe the system’s behaviour in infinity, or in practice, after a long time. Two methods of choosing the best strategy are presented depending on the type of problem we are facing. When the Decision Maker has the possibility to pre-select the states in which the enterprise will be found during their business, the expected prizes for each selected set of states are compared. If the states in which an enterprise can be found are fixed, and the Decision Maker has in every state the possibility to make certain decisions affecting the transition matrix (and also often on the matrix of awards), a maximization process algorithm is used to determine the best strategy. At the end is given a brief overview of the continuous Markov chains and it is shown how the theory developed on discrete Markov chains can be applied to a continuous case.