Purpose: To evaluate the effect of a reducing agent and plant-extract antioxidant on the bonding durability of a self-etch adhesive to normal and NaOCl-treated, smear-layer-deproteinized dentin. Materials and Methods: Flat smear-layer-covered dentin surfaces from 60 extracted human molars were prepared by removing the occlusal enamel. The teeth were divided into two groups with or without NaOCl-deproteinizing treatment for 30 s, and further divided into three subgroups as follows: no application of antioxidant, application of Accel (p-toluenesulfinic acid sodium salt solution) for 5 s, or application of rosmarinic acid solution for 5 s. All treated dentin surfaces were bonded with a two-step self-etch adhesive (Clearfil SE Bond) and restored with composite (Clearfil AP-X). The bonded teeth were sectioned into a hourglass-shaped sticks with a composite-dentin bonded interface area of 1.0 mm2. After storage in artificial saliva for 24 h or 1 year, the specimens were subjected to the microtensile bond strength test (n = 15). Data were statistically analyzed with three-way ANOVA, Tukey's post-hoc test, and the t-test (p 0.05). Results: Without an antioxidant, 1-year storage significantly reduced the bond strengths of the self-etch adhesive to normal and smear-layer-deproteinized dentin compared with those after 24-h storage (p 0.05). Application of Accel and rosmarinic acid restored the compromised initial bond strengths to smear-layer-deproteinized dentin (p 0.05), and prevented long-term deterioration of bond strengths to both normal and smear-layer-deproteinized dentin (p > 0.05). Conclusion: Application of Accel and rosmarinic acid improved bonding durability of the self-etch adhesive to both normal and smear-layer-deproteinized dentin. Schlagwörter: self-etch adhesive, dentin bonding durability, anti-oxidant/reducing agent, cross linker, MMP-inhibitor, smear-layer deproteinizing