The lack of passive adaptation in prosthesis over implant can overload implants, prosthetics components and bone tissue, leading to marginal bone loss and even osseointegration. Weldering is used frequently to improve the settlement of metal structures of prosthetic in general and prosthesis over implants. In addition, the settlement precision of the workpiece may be related to the care in the technique of manufacturing the same. The objective of this study was to evaluate, through optical and confocal Laser microscopy, the marginal adaptation of metal bars Branemark protocol type alloy nickel-chromium (NiCr) monoblock and welded, made from conventional waxing .For this, a master model was made to simulate four implants were adapted over which transferents and analogs that served to transfer the position of the implants for the cast model. Waxing obtained by the conventional method were cast in one-piece and separate parts (sectioned), the latter being subsequently welded using the conventional techniques blowtorch and TIG (Tungsten Inert Gas).The bar adjustment of readings in each implant were made in monoblock pieces (n = 5) and the severed pieces (n = 10) before and after different weldering techniques. By optical microscopy measurement method, measurements were performed in two predetermined regions (Vestibular and lingual), with three replications, totaling 6 readings per implant. However in the methodology of Confocal laser microscope, was done only on Vestibular. To compare the data obtained on the misfit between prosthetic/implant component, considering the pieces monoblock, the welded parts, the weldering techniques (torch and TIG), measurement methods and the position of the implants, tests were performed statistical where for each of the analyzes determined the application of Normality Test (Shapiro-Wilk test and Kolmogorov-Smirnov test (alpha = 5%)) to determine the distribution of the sample. For the sample of normal distribution was determined using the parametric test, analysis of variance, while the sample with non-normal distribution was determined using the nonparametric Kruskal-Wallis test. The results found were statistically significant for method and technique comparison, suggesting that the monoblock parts showed higher misfit than welded parts, and the parts welded by conventional technique showed no statistical difference in relation welded by TIG technique. Regarding the comparison of measurement methods, there was no statistically significant difference, where the Laser Confocal microscopy obtained the greatest misfit. The technical interaction and method there were no statistical differences. When comparing different sections and welded to the two measurement methods, in relation to Section factor, there was no significant difference where 3 sections have the lowest results and welded had equality with 1 section. The technical factor was no statistically significant difference, the conventional technique had the highest averages. The interaction and Technical Section had no significant...