Introduction: Underwater visual censuses are the basis of many studies on fish ecology, however, a series of limitations and errors influence the traditional visual estimation of fish richness and abundance. Video techniques have been proposed to mitigate such errors, but there are few studies that compare the effectiveness of both methods. Objective: To compare the estimates obtained through the traditional census and the video census of the fish community of two localities in the central Mexican Pacific. Methods: We studied the fish community of two bays of Huatulco, Oaxaca, Mexico. We established sampling points in each bay and applied a traditional census and a diver-operated video census. We used comparison tests and analysis of similarity tests to compare richness, abundance and diversity by locality; and permutation tests for the same parameters at each sampling point. Results: Both censuses provide similar estimates regarding the richness, abundance, and diversity by locality and by sampling points. There were no statistically significant differences between traditional census and a diver-operated video census in terms of richness, abundance, and diversity. Conclusions: Video census using the diver-operated video technique can be used as a complement or as an alternative to traditional census. Its use can provide a more complete assessment, increase data acquisition, and implement long-term monitoring programs in areas where there are economic limitations for its operation.