The quasielastic (QES) and the inelastic scattering (INS) observed in the neutron scattering spectra of polycrystalline Na 3Cr 2P 3O 12 (noted NaCr) from 0.8 to 673 K is due to spin fluctuations. The QES of NaCr and NaFe were analyzed within de Gennes' (DG) theory in a previous article. The present work is devoted to the INS of NaCr, it was measured from 0.8 to 673 K, at the time of flight IN6, IN5 and D7 (polarized neutrons) spectrometers at ILL. The Q dependence of the frequency maximum, of the half-width and of the amplitude of the INS have been investigated and discussed in the light of different models and theories such as the zero-field splitting (ZFS), the exchange coupled Cr 3+ dimers (ECCD) model and Lovesey and Meserve's (LM) theory. Although qualitative agreement can be obtained with the ECCD model, the LM model is more suitable for reproducing the experimental data. In particular the QES and the INS are treated simultaneously. For the last two models the average Jvalue ( ca 0.35 meV) which must be introduced is much larger than the value which was derived using the DG theory from the QES study of NaCr (0.083 meV), but of the same order as that found for NaFe in its paramagnetic phase (0.35 meV), this value being close to those found from low temperature measurements of magnon energies in NaFe and associated to the nearest Fe-Fe exchange couplings (0.2 meV), as if only these couplings were responsible for the paramagnetic fluctuations. For NaCr the situation is different since it is not possible to obtain consistent J values when one treats the QES and INS separately. Thus it remains difficult, both from a chemical point of view and with the J values reported above, to explain why for NaCr the INS persists to 673 K while for its Fe isomorph it disappears above 47 K. In any case the fact that NaCr does not order cannot be attributed to a small J value and must be found in the structural disorder associated to the Na + ions and to the presence of magnetic frustrations.