The goal of this study is to compare the misfit (>150µm) generated once the restoration, made by different techniques, is retained to a single conical implant. 15 internal connection implants (MIS C1 4?20x10mm) are embedded each one perpendicularly to an horizontal surface of the 1x1x2cm poliuretan resin model. The 15 samples obtained are divided in 5 groups depending on the framework process (n=3): 1/casting, 2/overcasting, 3/Ti-base, 4/milling and 5/laser sintering. The cobalt-chromium alloy frameworks are screw-retained to their respective implants to a 30-Ncm torque. Once it is retained the framework to the implant, the next step is to section the sample in half with a diamond saw and verify the correct fit with a stereomicroscope, measuring 4 distances in each side (A, B, C and D). Data is submitted to one-way analysis of variance (ANOVA). According to equality of variances, significant differences are found in A and B measures (p=0,000 in left side in both groups and, p=0,007 and p=0,001 in right side). In C and D, there are not statistical differences (p=0,586 and p=0,110 in left side and, p=0,101 and p=0,089 in right side). However, once it has realized ANOVA test, only C retains the hypothesis and accepts independence. More samples are needed to conclude reliable statements. However, what it is observed is that milled group presents the best marginal fit. Overcasted and Ti-Base abutments also have good results above casted ones, and, sintered groups has the lowest result. Although, all systems have gaps below 150 µm, so all of them are good options for prosthetic rehabilitation.