The aim of this study was to evaluate the erosion-preventive effect of different artificial saliva formulations and human saliva in vitro compared to human saliva in situ. In the in vitro experiment, bovine enamel and dentin specimens were stored in artificial saliva (4 different formulations, each n = 20), deionized water (n = 20) or human saliva (n = 6 enamel and dentin specimens/volunteer) for 120 min. In the in situ experiment, each of the 6 enamel and dentin specimens was worn intraorally by 10 volunteers for 120 min. The specimens were then eroded (HCl, pH 2.6, 60 s). Half of the specimens were subjected to microhardness analysis (enamel) and the determination of calcium release into the acid (enamel and dentin), while the other half were again placed in the respective medium or worn intraorally, respectively, for 120 min before a second erosion was performed. Knoop microhardness of enamel and the calcium release of enamel and dentin into the acid were again determined. Statistical analysis was conducted by two-way repeated-measures ANOVA or two-way ANOVA (α = 0.05). Enamel microhardness was not significantly different between all test groups after the first and the second erosive challenge, respectively. Enamel calcium loss was significantly lower in situ compared to the in vitro experiment, where there was no significant difference between all test groups. Dentin calcium loss was significantly lower than deionized water only after the first and than all except one artificial saliva after the second erosion. Under the conditions of this experiment, the use of artificial saliva formulations and human saliva in vitro does not reflect the intraoral situation in dental erosion experiments adequately.