Résumé : Abstract Enamel demineralisation can develop on smooth surfaces as an undesirable side effect during orthodontic treatment with fixed appliances. This study aimed to evaluate the ability of 38% silver diamine fluoride in remineralisation (as estimated by fluorescence gain) of artificial initial lesions in smooth surfaces of human enamel. The smooth surfaces of 50 human tooth samples were artificially demineralised and 45 samples were allocated randomly into three groups receiving a single treatment with a varnish: group I: Riva Star (silver diamine fluoride, SDF), group II: Bifluorid 12 (NaF, CaF 2 ), and group III: Cervitec F (CHX, CPC, NH 4 F). Five samples were assigned as a negative control group without treatment. All samples were exposed to pH-cycling for 28 days. Fluorescence behavior was measured using Quantitative light-induced fluorescence before and after demineralisation and up to four weeks on a weekly basis. Analysis of variance (ANOVA) with Tukey–Kramer post-hoc tests and repeated measures ANOVA were used for statistical evaluation (α = 0.05). After demineralisation, all samples showed mean ΔF of − 16.22% ± 4.35, without significance differences between the fluorescence behaviour of the samples ( p  = 0.251). After 28 days group comparison showed a statistically significant difference ( p  = 0.034) for ΔF values: the lowest fluorescence values were found in group I (SDF, mean ΔF − 16.47 ± 6.08) with a significant difference compared to group III (Cervitec F, mean ΔF − 11.71 ± 4.83). In group II (Bifluorid 12) mean ΔF value was − 15.55 ± 2.15) without statistically significant differences to groups I and III. The fluorescence behaviour of SDF varnish on smooth surfaces with artificial initial enamel lesions was significantly lower compared to Cervitec F varnish after short time use.