The compressor is an important auxiliary for fuel-cell (FC) operation. Growing fuel-cell system efficiency involves an optimal fuel cell energy management and the air management is a key issue. Thus, a good modelling for static and dynamic operation of all components of the FC system, and in particular of the compressor, is required. The difficulties, due to a lack of information about the performance of compressors, demand predictive and modern approximation methods to be used for compressor modelling. To overcome these issues, the paper proposes and presents a moving least squares (MLS) algorithm for obtaining a surrogate model of the centrifugal compressor. The experimental data provided by manufacturers are used for this task. The results can be used for the development of an off-design model or the overall dynamic simulation of the behaviour of a FC system.