Objective The current study aimed to investigate the bonding properties of a novel low-shrinkage resin adhesive containing expanding monomer and epoxy resin monomer after thermal cycling aging treatment. Methods Expanding monomer of 3,9-diethyl-3,9-dimethylol-1,5,7,11-tetraoxaspiro-[5,5] undecane (DDTU) as an anti-shrinkage additive and unsaturated epoxy monomer of diallyl bisphenol A diglycidyl ether (DBDE) as a coupling agent were synthesized. A blend of DDTU and DBDE at a mass ratio of 1∶1, referred to as “UE”, was added into the resin matrix at the mass fraction of 20% to prepare a novel low-shrinkage resin adhesive.Then, the methacrylate resin adhesive without UE was used as the blank control group, and a commercial resin adhesive system was selected as the commercial control group. Moreover, the resin-dentin bonding and micro-leakage testing specimens were prepared for the thermal cycling aging treatment. The bonding strength was tested, the fracture modes were calculated, the bonding fracture surface was observed by scanning electron microscope (SEM), and the dye penetration was used to evaluate the tooth-restoration marginal interface micro-leakage. All the data were analyzed statistically. Results After aging, the dentin bonding strength of the experimental group was (19.20±1.03) MPa without a significant decrease (P>0.05), that of the blank control group was (11.22±1.48) MPa with a significant decrease (P<0.05) and that of the commercial control group was (19.16±1.68) MPa without a significant decrease (P>0.05). The interface fracture was observed as the main fracture mode in each group after thermal cycling by SEM. The fractured bonding surfaces of the experimental group often occurred on the top of the hybrid layer, whereas those of the blank and commercial control groups mostly occurred on the bottom of the hybrid layer. Micro-leakage rating counts of specimens before and after thermal cycling were as follows: the experimental group was primarily 0 grade, thereby indicating that a relatively ideal marginal sealing effect could be achieved (P>0.05); meanwhile, the blank control group was primarily 1 grade, and the penetration depth of dye significantly increased after thermal cycling (P<0.05); the commercial control group was primarily 0 grade without statistical difference before and after thermal cycling (P>0.05), while a significant difference was observed between the commercial control group and experimental group after thermal cycling (P<0.05). Conclusion The novel low-shrinkage resin adhesive containing 20%UE exhibited excellent bonding properties even after thermal cycling aging treatment, thereby showing a promising prospect for dental application.