The fwhm tells you about the homogeneous/inhomogeneous broadening of the natural linewidth of the transition, which an be relater=d to crystallinity disorder. First check the lineshape- a Lorenztian is an ideal lineshape, and deviation to a Gaussian indicated disorder. A PLspectrum may also have phonon sidebends, however, in which case you need to deconvolute the peaks.

As an addition to the answer of Hugh J Byrne, the inhomogeneity might be caused by band gap fluctuations, which can be due to chemical fluctuations or electrostatic potential variations. Latter one can be also related to the degree of compensation, i.e. the ratio of donors to acceptors. If you have locally a high density of donors or acceptors, which cannot be screened, then you get band gap fluctuations leading to a broadening of your PL peak. Hence, the FWHM of a peak also reveals insights to the doping of your investigated material.

Please read the attached articles which have explained how PL intensity, FWHM, Peak position affect properties of halide perovskite. Another article explain about line shape and width of PL spectra.