In many engineering fields it is beneficial to obtain information about the force acting on a dynamical system. As measurement of this force is often difficult or impossible a technique that identifies this force in an alternative manner is desired. A wide variety of methods is available in literature. Obtaining this force in the frequency domain is done often. However, in certain cases where the input force is non-stationary a frequency domain technique does not suffice. This thesis therefore focuses on obtaining a reliable force identification method in the time domain. The force identification problem can be seen as an ’inverse problem’ to which a simple analytical solution is not trivial. A more advanced method is required. Methods found in literature can be grouped into three categories which fundamentally differ in the way the dynamics is modelled. Deterministic force identification methods are defined as methods where the dynamics is modelled deterministically. Whenever a methods uses a stochastic model it is considered a stochastic force identification method. A third group of force identification methods uses artificial intelligence to obtain a model of the system when no model is available. In this thesis it is assumed a model of the system dynamics is available and therefore artificial intelligence methods for force identification are not considered. Deterministic force identification method including regularization methods, recursive methods and iterative methods are compared to stochastic methods which are all based on the Kalman filter. The most relevant methods are evaluated using simulated data of a single and multiple degree of freedom dynamical system and measured performed on an aluminium structure. It was concluded that the Least Mean Square Adaptive Algorithm outperforms the Joint Input-State Estimator with Artificial Displacement Measurements in identifying forces acting on the simulated single and multiple degree of freedom system as well as the forces acting on the aluminium structure.