The aim of this study was to evaluate gap formation after using two universal adhesives, Peak Universal Bond (P), which contains chlorhexidine and Adhese Universal (A), which contains 10‐MDP monomer, and to compare them with two gold standard adhesives. The adhesives were tested in different etching modes, etch‐and‐rinse (ER) or self‐etch (SE), and under thermomechanical loading (TML). Two identical Class V cavities were prepared in each of 30 human third molar teeth, and resin‐composite restorations were made. Before and after TML, replicas of the restorations were prepared. Gap formation and their frequency according to the modified classification of Blunck and Zaslansky and the width of the maximum marginal gap (MG) were separately evaluated in enamel and dentin. Gap frequency did not differ significantly only in the enamel before TML, despite the use of different adhesives or etching modes. At both timepoints, the highest MG values were obtained with the Peak Universal Bond self‐etch (PSE) group in enamel (6.2 μm before TML, 12.2 μm after TML) and the Peak Universal Bond etch‐and‐rinse (PER) group in dentine (21.3 μm before TML, 35.5 μm after TML). After TML, there was a significant increase of gap frequency for Adhese universal self‐etch (ASE) group and MG values for PER, PSE, ASE groups at the resin‐enamel interface, and gap frequency and MG values for all groups at the resin‐dentin interface. Universal adhesive with 10‐MDP monomer have beneficial effect on the resin‐dentin interface, while chlorhexidine may reduce the resistance of the universal adhesive system to oral cavity conditions.