Stock exchanges are modeled as nonlinear closed-loop systems wherethe plant dynamics is defined by known stock market regulations and the actions ofagents are based on their beliefs and behavior. The decision of the agents may containa random element, thus we get a nonlinear stochastic feedback system. The marketis in equilibrium when the actions of the agents reinforce their beliefs on the pricedynamics. Assuming that linear predictors are used for prediction of the price process,a stochastic approximation procedure for finding market equilibrium is described.The proposed procedure is analyzed using the theory of Benveniste et al. (Adaptivealgorithms and stochastic approximations. Springer, Berlin, 1990).Asimulation resultis also presented.