This study evaluated the influence of a soft-start light-curing exposure on polymerization shrinkage stress and marginal integrity of adhesive restorations. Six resin-based composites (Pertac II, Tetric Ceram, Definite, Surefil, Solitaire, and Visio-Molar) were adhesively bonded to a cylindrical cavity (n = 9 per material/light) in a photoelastic material. Visible light-curing was applied using either the standard polymerization mode (800 mW/cm2 exposure duration 40 s) of the curing light (Elipar TriLight, 3M ESPE) or the exponential mode from the same device (ramp-curing: 150 mW/cm2 to 800 mW/cm2 within the first 15 s of a total curing time of 40 s). Polymerization stress was calculated at 5 minutes, 1 hour, and 24 hours postexposure from the second-order isochromatic curves obtained from photoelastic images (Matrox-Inspector). Two standardized Class V preparations were made each on the facial and lingual surfaces of 80 extracted human molars and premolars. Resin restorative systems (Pertac II/EBS Multi, Tetric Ceram/Syntac, Definite/Etch&Prime 3.0, and Surefil/Prime & Bond 2.1) were exposed using both light exposure modes (n = 20). Marginal dye penetration (2% methylene blue) was investigated separately for enamel and cementum margins after thermocycling. To obtain information on equivalent depth of cure, relative surface hardness measurements were performed on resin samples of the same material at the top surface and at 1.5 mm and 3.0 mm thickness (Zwick 3212, 10 N). A significant (p < .01) reduction in polymerization stress of 7.1% for Pertac II, 4.1% for Tetric Ceram, 3.6% for Definite, 3.7% for Surefil, and 6.2% for Solitaire was observed when using the exponential mode as opposed to the standard. A significant (p = .04) reduction of marginal dye penetration was found only for Pertac II/EBS Multi at the cementum margins when the soft-start polymerization was used. For the sample thickness of 3 mm, a significant higher relative bottom to top surface ratio in favor of the standard exposure mode was found (p = .001). Depending on the restorative material, soft-start polymerization may lead to a significant reduction in marginal microleakage of adhesive Class V restorations. This effect might be attributable to a significantly lower polymerization stress, as seen from photoelastic analysis, and/or a decrease in the degree of conversion, as deducted from surface hardness ratios. However, the effect of soft-start curing mode depends on the material itself, with the most effective response from hybrid resin-based composites.