Abstract The epifluorescence technique was used to quantify the total bacterial population in two dams (Neusa and Prado) and in a natural lagoon (Fúquene), which have different geographical locations and physicochemical characteristics. Bacterial abundance was similar in the three water‐bodies (8.4 × 10 8 cells L –1 ), but the density of plateable heterotrophs was always lower (1.4 × 10 6 CFU L –1 ) when obtained by the plate count method. Results obtained during validation of the epifluorescence technique show: (i) that there are no statistically significant differences between counts of samples fixed in the field and unfixed samples; (ii) that the median count of samples fixed and observed periodically over a period of 8 months was similar; (iii) that plate counts underestimate the population; and (iv) that epifluorescence counts were similar in the three water‐bodies studied, despite their different ecological conditions.