In this paper, the nonlinear friction characteristic of a custom made symmetrical linear hydraulic actuator is investigated using the Extended Kalman Filter (EKF). A new and very accurate characterization of friction is made by using a quadratic function of the piston velocity. Further to this proposed empirical friction model, the EKF is used to estimate the function coefficients. In this paper, an iterative approach is used to maintain system observability and render the estimation process more reliable. The study is conducted in simulation and by using measured experimental data. The estimated states and parameters by the EKF are found to be convergent to their known values in simulation and, further to experimental results, unique and repeatable. In addition, changes in the friction characteristics, which can occur in the physical system due to wear in the piston seals or degradation in the oil properties, are detected and accurately estimated by the EKF in simulation. This study presents an accurate nonlinear model for the representation of friction in a hydraulic actuator. It paves the way for the implementation of strategies for early fault detection in hydraulic systems.