Fluorescence indices are particularly well adapted to characterize dissolved organic matter. In environmental studies, the three most often used fluorescence indices are the Fluorescence index (FI), the Humification index (HIX) and the Biological index (BIX). The present paper aims to discuss 4 limitations in the actual use of fluorescence indices: (i) difference between indices values calculated from interpolated and non-interpolated excitation wavelengths; (ii) misfit between the areas and wavelengths from the index definition (HIX, FI, BIX) and the samples (peaks location and curve shape) related to environmental context (e.g. pH, living conditions of organic matter decomposers); (iii) the potential to rethink the 2D-based indices thanks to 3D matrices; (iv) the incidence of recent improvement in organic matter degradation knowledge on the fluorescence indices interpretation. These limitations are illustrated with different types of water samples coming from three different open access datasets: karst groundwaters from Serene et al. (2024), marine waters from Drozdova et al. (2019) and surface water from Sgroi et al. (2019). These results draw the attention of fluorescence indices users to unsupervised calculation and interpretation of 2D-based fluorescence indices. It also greatly encourages the rethinking of fluorescence indices using EEM (Excitation Emission Matrices) in order to adapt the calculation to each purpose and study context and to limit the generalisation of indices interpretation to organic matter structure.