Abstract : The problem of the optimal use of private information is omnipresent on the financial markets (Insider trading, problems of default, etc). To analyze such a problem properly, the interactions between agents are to be considered strategically : the information conveyed by the prices fixed by an informed agent influences the future behavior of the asset price. This opportunity of influencing price is generally not considered by the classical finance theory. Game theory is the natural framework to analyze strategically these interactions. The main aim of this thesis is precisely to develop financial theory based on game theory. De Meyer and Moussa Saley, in "On the origin of Brownian Motion in finance", model the interactions between two asymmetrically informed market makers by a zero-sum repeated game with lack of information on one side. In particular, this study shows that the Brownian motion, often used in finances to describe the price dynamics, has partially a strategic origin : it is introduced by the informed agents in order to take maximal benefit from their private information. This thesis deals with several generalizations of this model about : Influence of the price grid, the bilateral asymmetry of information and the diffusion process of the information.