It is shown that the task of ribonucleic acid (RNA) folding can be represented as a task of making optimal decisions in game theory. The difference is, however, the impossibility of modelling (calculating) all possible RNA chain conformations (states) because the task is NP-complete. Due to that, the goal-oriented search has to be made, i.e., which states should be calculated and which should not. For that purpose, the X-tuning method is suggested that is based on the most close to the initial state actions (moves and turns) and thus significantly reduces the searching. Despite that this method does not ensure obtaining of the global optimum, it enables obtaining a group of acceptable decisions, which is often enough in practice. It is explained that the X-tuning method can be applied both in games against enemy, and in the task of RNA folding (a game against nature).