Background: Performing an appropriate Class II direct restoration is a great challenge. The correct selection of the matrix system and the elements used for its stabilization will have a great impact on the result. The aim of this study is to show a new digital method for a predictable selection of the wedge and compare it with the conventional method. Methods: Sixty patients were randomly divided into two groups. In Group 0, the wedge was selected intraoperatively by visual examination, while in Group 1 the wedge was selected preoperatively through the measurement of the cervical embrasure on the bite-wing radiography. The number of wedges used, modifications to them, and tissue damage were registered, along with the quality of the proximal contact and the marginal adaptation. Results: Student’s t-test revealed a statistical difference between the number of wedges used, which was higher in Group 0 (p < 0.05). Pearson Chi-square test showed that tissue damage was statistically higher in Group 0 (p < 0.05), while there was no statistically significant difference between groups in wedge modifications (p > 0.05). Group 1 revealed a higher frequency of satisfactory proximal contact and marginal adaptation (p < 0.05). Conclusions: This new technique reduces wedges waste and tissue damage and provides an adequate interproximal anatomy.