The present study evaluated the effect of fluoride varnishes containing nano-sized sodium trimetaphosphate (TMP) on the remineralization of artificial caries lesions and erosion of dental enamel in vitro. As well as the pattern of F and TMP release from these formulations. The varnishes tested were: Placebo (without fluoride or TMP, negative control), 2.5% NaF, 5% NaF (positive control), 5% NaF + 5% TMP micrometric, 5% NaF + 2.5% nano-sized TMP, 5% NaF + 5% nano-sized TMP, besides a commercial formulation (Duraphat), hereafter referred to as PLA, 2.5%F, 5%F, 5%F+5%TMPmicro, 5%F+2.5%TMPnano, 5%F+5%TMPnano and Duraphat, respectively. In the first phase, artificial caries lesions were induced on bovine enamel blocks (n=168), which were selected by surface hardness (SH). Blocks received a single application of the aforementioned varnishes, remaining in contact for 6 h. Half of the blocks (n=12/group) were used to determine the concentration of CaF2 and FA fluoride formed on the specimens after the treatment with the varnishes. The other half was subjected to a pH cycling model (6 days). The blocks were analyzed for the percentage of SH recovery (%SHR), cross-sectional hardness (ΔKHN), CaF2 and FA retained after pH cycling. The varnishes containing TMP promoted significantly higher %SHR compared to 5%F, without significant differences among the TMP-containing varnishes. A similar pattern was observed for ΔKHN, although the values obtained for 5%F+5%TMPnano were 25% lower than those obtained for 5%F+5%TMPmicro. The highest CaF2 concentrations were promoted by 5%F, 5%F+5%TMPmicro and Duraphat varnishes. The second phase evaluated the protective effect of PLA, 5%F, 5%F+5%TMPmicro, 5%F+2.5%TMPnano and 5%F+5%TMPnano varnishes on initial erosion of bovine enamel. Specimens (n=8), selected by SH, received a single application of the varnishes, remaining in contact for 6 h. Varnishes were then removed and the blocks, submitted to four individual erosive challenges (1 minute, citric acid, 0.75%, pH = 3.5, under stirring) and analyzed by SH after each challenge. Overall, the highest percentage of SH change was observed for PLA, followed by 5%F, 5%F+5%TMPmicro and both varnishes containing TMPnano, with no significant differences between 5%F+2.5%TMPnano, 5%F+5%TMPnano. Finally, the third phase evaluated the pattern of fluoride and phosphate release from PLA, 2.5%F, 5%F, 5%F+5%TMPmicro, 5%F+2.5%TMPnano, 5%F+5%TMPnano and Duraphat varnishes over 24 h, in a pH cycling model. The varnishes were applied on polyester sheets (n=8/group), which were alternately immersed in remineralizing and demineralizing solutions at 30, 60, 90, 120, 180, 240, 300, 360, 420, 540, 600, 720 , 780, 960, 1200 and 1440 min after the first immersion. The solutions were analyzed for fluoride and phosphate concentrations. The varnishes containing TMP promoted an increasing, exponential pattern for the cumulative fluoride release up to 6 hours, reaching a plateau afterwards. In addition, higher quantities of fluoride were released when varnishes were immersed in the demineralizing solution. In general, varnishes containing TMPnano released significantly higher amount of fluoride compared to TMPmicro. The results of the 1st phase allow to conclude that the addition of TMP to fluoride varnishes significantly increases its remineralizing potential in artificial caries lesions, with an additional effect with the use of TMPnano in relation to TMPmicro, this increment was not statistically significant. As for the effect against erosive challenges (2nd phase), a similar pattern was observed, with the highest protective effect observed for both varnishes containing TMPnano, which was significantly higher than the other groups. The higher effect of varnishes supplemented with TMP seems to be related to the higher fluoride release from these products (3rd phase), especially TMPmicro, associated to the constant release of TMP from these formulations.