Electromagnetic actuators (EMA), which incorporate solenoids, are increasingly becoming the actuator of choice in industry lately, due to their ruggedness, low cost, and relative ease of control. Latest applications of solenoid based EMA’s include Electromagnetic Valve Actuation (EMV) systems. This application presents challenges that require the improvement of the dynamic characteristics of the EMA. Some of these problems include, but are not limited to, quiet operation, reduced bounce, less energy consumption, trajectory shaping with a minimum number of measurements, and high actuation speeds. These demands, coupled with the nonlinear dynamics of the EMA, make the use of classical control strategies a less attractive option. A possible attempt to arrive at intermediate solutions to these problems should include some amount of model based robust control strategy. This includes the development of an accurate but simple control based model and a robust digital control strategy. In this study a basic nonlinear model for a solenoidal EMA will be developed, and validated, which will include bounce, leakage inductance and temperature effects. The model is formulated for the linear legion (region before saturation) of the actuator dynamics, but validation will include operation in the saturation region as well. This effectively means that a nonlinear model will be developed that is simple but accurate enough for control, neglecting hysteresis and magnetic saturation. Next, an EMV will be designed and built. A nonlinear model for the EMV will be developed and validated. This model will include secondary nonlinearities like saturation, hysteresis, mutual inductance and bounce. In this study a variable that is easier and cheaper to measure, current, will be measured and the information of the position and velocity variables will be estimated from this measurement. The position estimate will be used for control. This is called Sensorless Control. The control objective is to reduce impact noise and seating velocity. The sliding mode methodology will be used here since it is nonlinear, robust to uncertainties, and easier to design and implement. The estimation and control algorithms will be validated in simulation and experimentally for the EMA and EMV, respectively.