This paper deals with a Mean Value Engine Model (MVEM) approach to dynamically estimating engine load torque. The load torque estimator is made using parameter estimation and an extended Kalman filter. The load torque estimator can be used to simplify the control logic of modern engines as it can detect load changes over time thereby accounting for wear and also differing accessory power consumption. The model can be implemented using a state equation for the crank shaft speed, and a state estimating a constant describing the loading level.