The aim of this study was to evaluate the efficacy of a pouring technique for implant-supported prostheses impressions. A metallic matrix (control group) with two implants positioned at 90 and 65 degrees was fabricated. The matrix was submitted to the direct transfer impression technique. In group CP (conventional pouring - n = 10), casts were obtained by the conventional pouring technique. In group EP (experimental pouring - n = 10), the analogs were embraced with latex tubes before the first pouring and then submitted to a second pouring. Vertical misfit and implants/analogs inclinations were evaluated. Data were analyzed by analysis of variance and Tukey's test (p < .05). Results demonstrated significant difference (p < .05) between control and experimental groups for misfit measurement in perpendicular implant/analog and between control group and group EP in leaning implant/analog. Considering inclination, there were significant differences (p < .05) between control and experimental groups for leaning analogs. Independently of the pouring technique, perpendicular implants produced more accurate casts.