Objectives: The present study aimed primarily to explore the cone angle of in vitro CAD/CAM conic crowns in pursuit of reversible retention in fixed implant-supported prostheses as an alternative to cement- or screw-retained prostheses. It further sought to develop predictive models for clinical applications. Materials and Methods: Eight sub-sets of CAD/CAM specimens consisting of an abutment and a coping with cone angles that varied from 1 to 8°. The retention surfaces of the elements were designed to provide very close contact between them. Five specimens were fabricated per angle and the retention strength was measured (in newton) five times for each specimen on a Zwick/Roell testing frame. The findings were analysed statistically with classical exploratory and descriptive tests of the quantitative variables, as well as with analysis of variance tests: Anova followed by Tukey multiple contrasting; and estimates models, calculating R2 parameters and goodness of fit. Results: The mean retention strength values found, in descending order of cone angle, were as follows: 8°: 21.02 N; 7°: 23.16, 28 and 36.40 N; 6°: 40.46 N; 5°: 66.36 N; 4°: 61.23 and 76.12 N; 3°: 93.44, 103.21 and 112.04 N; 2°: 154.20 N and 1°: 204.74, 261 and 293.40 N. These data described a curve from which two predictive models were developed to find retention strength from the cone angle used and the cone angle that would be needed to deliver a given retention strength. Conclusions: On the grounds of these findings and subject to the limitations of in vitro studies, the conclusions drawn were: in CAD/CAM conic crowns, the smaller the cone angle, the greater the retention strength; predictive models can be developed for both cone angle and retention strength; and lastly, this line of research opens the door to a wide field of future studies. Palabras clave: Implant-supported, Double Crowns, Adherence