Fisheries managers rely on independent programs to collect data to design and inform management plans and decisions. These programs aim to implement cost-effective, robust monitoring approaches to assess spatiotemporal changes in fish structure and composition using different methods, which could lead to different conclusions. We compared four methods: two types of underwater visual censuses (UVC) (i.e. transect and roving), remote underwater video (RUV), and fishing using handline (HL) to characterize differences, identify potential complementarities, and estimate relative precision among them. All methods were simultaneously performed on 13 sites in Fajardo, eastern Puerto Rico for a total of 26 visits, and nine sites in La Parguera Natural Reserve (LPNR), western Puerto Rico for a total of 43 visits. Multivariate analysis revealed significant dissimilarities between all methods, with higher species diversity, abundances, and biomass observed in UVC and RUV. Patterns of spatial and temporal variation detected with UVC and RUV were significantly correlated among them, but not with HL. Univariate analysis for commercially important species showed that UVC and RUV methods were more precise and time efficient than HL. Complementarity between UVC and RUV was observed with the latter increasing the diversity observed during surveys. This study showed that UVC and RUV were different, however, they can be used complementarily they were significantly correlated. HL should be used when additional information (e.g. sex and age) is needed. A multi-gear approach with increased use in RUVs is recommended depending on the aims of the study to precisely assess a broader diversity.