Optimizing the heating properties of magnetic nanoparticles is of great importance for hyperthermia applications. Recent experimental results show that core/shell nanoparticles could give an increased specific absorption rate (SAR) compared to the magnetic oxide nanoparticles currently used. We have developed a modified phenomenological model based on the linear Néel-Brown relaxation model to calculate the SAR due to susceptibility losses in complex nanoparticles with ferromagnetic (FM) core/ferrimagnetic (FiM) shell morphology. We use the Monte Carlo (MC) simulation technique with the implementation of the Metropolis algorithm to investigate the effect of size and shape on the magnetisation behaviour of complex ferromagnetic/ferrimagnetic nanoparticles covered by a surfactant layer. The findings of our simulations are used as an input in our modified model for the calculation of the SAR. Our calculations show that for all the sizes and shapes the complex FM/FiM nanoparticles give higher SAR values than the pure ferrimagnetic ones due to their higher core saturation magnetisation. For all sizes the nanoparticles with the truncated cuboctahedral shape give the highest SAR values and the cubic ones the lowest ones. The decrease in the surfactant thickness results in an increase of the SAR values. Our results have the same characteristics as the available experimental data from Fe/Fe3O4 nanoparticles, confirming that the complex nanoparticles with core/shell morphology can optimise the heating properties for hyperthermia.