Physically motivated models of electromechanical motion systems are required in several applications related to control design. However, the effort of modelling is high and automatic modelling would be appealing. The intuitive approach to select the model with the best fit has the shortcoming that the chosen model may be one with high complexity in which some of the parameters are not identiifable or uncertain. Also, ambiguities in selecting the model structure would lead to false conclusions. This paper proposes a strategy for frequency domain model selection ensuring practical identifiability. Also, the paper describes distinguishability analysis of candidate models utilising transfer function coecients and Markov parameters. Model selection and distinguishability analysis are applied to a class of models as they are commonly used to describe servo control systems. It is shown in experiments on an industrial stacker crane that model selection works with little user interaction, except from defining normalised hyperparameters.