This study evaluated the effects of thermocycling on the microtensile bond strength (microTBS) of one- and two-step self-etch adhesives (SEAs) to sclerotic dentin. Two adhesives, Clearfil S3 Bond (S3), a one-step self-etch adhesive (1-SEA), and Clearfil SE Bond (SE), a two-step self-etch adhesive (2-SEA), were applied on cervical lesions in human premolars with sclerotic or normal dentin. After adhesive application, the lesions were restored and built up using a resin composite (Clearfil AP-X). After 24 hours in water storage, the restored teeth were sectioned into 0.7 x 0.7 mm composite-dentin beams. The beams were then aged with 0, 5,000 or 10,000 thermocycles. The use of two adhesives, two substrate types and three thermocycling regimens yielded 12 experimental groups of 14-19 beams each. The beams were subsequently subjected to microTBS testing at a crosshead speed of 1 mm/minute and statistical analyses were computed with three-way ANOVA and Tukey's post hoc test at p < 0.05. Three-way ANOVA showed statistically significant effects on bonding effectiveness by lesion type, adhesive system, thermocycling or combinations of the adhesive system and thermocycling (p < 0.05). With sclerotic dentin, although S3 and SE provided comparable microTBS after 24 hours of water storage, S3 showed significantly lower microTBS than SE after thermocycling (p < 0.05). Regardless of lesion type, the microTBS for S3 decreased significantly after 5,000 or 10,000 thermocycles, while the microTBS for SE showed a significant decrease only after 10,000 thermocycles. Regardless of the extent of thermocycling, the microTBS values for either SE or S3 bonded to sclerotic dentin were significantly lower than to normal dentin (p < 0.05). The results suggested that thermocycling had a significant negative effect on the bond strength of the two SEAs tested. In contrast to 2-SEA, 1-SEA might not be a good choice for sclerotic dentin when seeking durability of the resin-dentin bond.