The aims of this study were, first, to measure shrinkage induced by different methods of pulse-delay light curing and, second, to verify their influence on the marginal adaptation of class V restorations in enamel and dentin. Eight groups, comprising seven groups (n = 6) with different pulse-delay parameters and a control group, were compared for dynamic linear displacement and force by using a fine hybrid composite. Based on these results, the pulse-delay curing procedure with the lowest shrinkage force was chosen and tested against the control group with respect to marginal adaptation in class V restorations (n = 8) before and after simultaneous thermal and mechanical loading. Statistically significant differences between groups were found for both shrinkage properties tested, with one pulse-delay group giving the lowest overall shrinkage values. However, the percentages of 'continuous margin' of this group, and of the control before and after loading, were not significantly different in dentin, whereas a significantly lower percentage of 'continuous margin' was detected in the pulse-delay group in enamel after loading.