ObjectivesTo evaluate in vitro the adhesion system of direct and indirect composite resin restorations in enamel and dentin, in cavities made in the cervical region using microtensile bond strength (μTBS) and to observe the most recurrent types of failure in the different groups. MethodsStandardized cavities were made in the line of the cementoenamel junction (CEJ) in 36 selected human premolars, which were randomly divided into 2 groups. In the Indirect Restorations (IR) group, indirect restorations were cemented in composite resin, using a universal adhesive and dual resin cement, performing selective enamel etching (SEE). In the Direct Restorations (DR) group, direct restorations were made in composite resin, performing SEE and bonding with the universal adhesive. Each specimen was sliced into sections measuring approximately 1 mm2, with enamel and dentin adhesion divided into subgroups DRe, DRd, IRe, and IRd. The sections were subjected to the μTBS test to assess bond strength in both enamel and dentin, comparing direct and indirect restorations in the cervical region. The types of failure in each group were also observed. ResultsThere was no statistically significant difference in μTBS between the DRe and IRe groups (p = 0.314) and also between the DRd and IRd groups (p = 0.145). There was a predominance of adhesive failures. ConclusionsThe bond strength of composite resin restorations was found to be equivalent in enamel and dentin, whether applied directly or indirectly. There was a predominance of adhesive fractures followed by cohesive resin fractures. Clinical significanceThe choice between direct and indirect restorations in the cervical region may consider patient preferences, but with caution, as clinical factors such as operator skill, specific tooth conditions and case complexity are crucial, and both techniques exhibit comparable bond strength.