In dairy manufacturing plants, biofouling of separation membrane represents a serious quality issue. Besides hydrodynamic conditions, the adhesion of pioneer bacteria and the formation of biofilms during filtration of dairy fluids could be influenced by membrane material properties. Consequently, the objective of this study was to characterize the impact of 3 different membrane materials (polyethersulfone [PES], polyvinylidene fluoride [PVDF], polyacrylonitrile [PAN]) on the diversity of early bacterial communities formed on membranes after ultrafiltration (UF) of dairy fluids. A laboratory-scale crossflow filtration system equipped with parallel modules, each containing with a different membrane of 42 cm², was used for UF of pasteurized skim milk and cheese whey. The UF system was operated at 50°C during 5h- and 20h-periods in the concentration mode with the feed maintained at 10°C between passages in the UF system. Membranes were cleaned with an alkaline solution prior and following each UF experiments. The bacterial diversity was assessed on cleaned membrane coupons and in filtered fluids after UF of 5 h and 20 h through a metabarcoding approach targeting the 16S rRNA gene. The bacteria numeration in samples was also estimated using qPCR targeting the same gene target. Bacterial genus ratios within the biofilms were found dependent of the composition of the membrane material used during UF of milk and whey. Interestingly, the qPCR quantification revealed a similar number of bacteria for each condition (P > 0.05). According to a PERMANOVA analysis, the diversity observed on membranes was dependent of the nature of the filtered fluids and the filtration duration, explaining respectively 53.24% and 21.75% (P = 0.002) of the variances among bacterial communities. Consequently, this study suggests that the membrane material may affect the biofilm formation on UF membranes, but other operational parameters such as the duration of filtration between cleaning cycles should be prioritized to control the biofouling issue.