Induction welding involves generating heat by applying an oscillating magnetic field, which produces eddy currents and Joule losses in an electrically-conductive material or hysteresis losses in a magnetic material. Most applications rely on eddy currents generation as composites are often made of electrically-conductive carbon fibres. However, in other applications, heat can be produced by a magnetic susceptor located at the weld interface of the parts to be joined. Composite films of magnetic particles dispersed in a thermoplastic matrix can serve as magnetic susceptors. Magnetic particles selection relies on various parameters that must be thoroughly defined beforehand. Firstly, the applied magnetic field amplitude and frequency is calculated, based on the generated current and the induction coil geometry. Secondly, the thermoplastic matrix is characterized, mainly with DSC measurements, to define its processing window. Finally, the magnetic properties of the particles are measured - for instance using a vibrating sample magnetometer (VSM) - to obtain the hysteresis curve for the applied field. The enclosed surface area of the hysteresis curve (i.e. absorbed energy density) is critical, as low hysteresis materials (i.e. soft magnets) will not dissipate enough heat, while high hysteresis materials (i.e. hard magnets) cannot be fully exploited as the saturation hysteresis is not reached within the used field amplitude. A methodology to approximate the hysteresis enclosed surface area with limited data is proposed, helping to anticipate the heating rate of a susceptor candidate material. Based on these parameters, the theoretical heating rates of three magnetic susceptor materials (magnetic particles of iron, nickel and magnetite) for induction welding are calculated. They are verified experimentally by comparing with the hysteresis analysis and by measuring the temperature evolution of samples made of polypropylene containing the magnetic particles