In this paper, the problem of finite-time model reference adaptive state tracking control and the weighted H∞ performance are studied with state unmeasured and actuator saturation. Compared with the existing literature, the state directly measurable problem is relaxed through a state observer. A model reference based adaptive controller is designed to address the state tracking error. Considering the disdifferentiability of the state for the considered system when the controller is saturated, an auxiliary design function is proposed to solve this problem. By utilizing Lyapunov stability theorem, finite-time tracking error bounded and the finite-time weighted H∞ performance are investigated. Moreover, the conservatism of the considered system is investigated by defining the different saturation rates. Finally, examples are provided as a means of illustrating the effective design method.