This paper presents a new method to estimate vehicle velocity and tire-road friction forces using an extended Kalman filter (EKF). A stochastic dynamic model is considered for tire-road friction forces and the vehicle motion equations are augmented to include friction force models. The state vector of this system consists of the vehicle longitudinal and lateral velocties, yaw rate, wheel angular velocities and friction forces. A square root algorithm is used for EKF implementation. The main features of this method are numerical stability and computational efficiency. The effectiveness of the proposed method is demonstrated by simulation.