This thesis presents a scheme for the parameter estimation and stabilization of an inverted pendulum system. First, a stable feedback loop is constructed for the unstable nonlinear system and then the parameters of its linearized model are estimated using input-output data, a real-coded genetic algorithms (RCGA) and the model adjustment technique. Parameter estimation does searching linear model from nonlinear model. A PI-type LQ control scheme is designed on the basis of estimated linear model such that the output can track a step reference input under constant disturbance environment. The performance of the proposed scheme is demonstrated through a set of simulation and experiment.