Aim: To introduce a modification of the reline impression technique (MRIT), and compare the dimensional changes of impressions obtained by MRIT and by conventional reline impression technique (CRIT). Methods: An acrylic resin tablet was milled by a CAD-CAM system to simulate three abutments (A, B and C) with different distances among them. The abutments were molded using both impression techniques. For MRIT, before completing the putty silicone polymerization, the relieve procedure was made by compression and it was immediately repositioned to complete the polymerization. Impressions were stored dry at room temperature for different periods (immediately, 1 h, 2 days and 7 days). The distances were obtained by scanning. The differences between the impressions and their respective matrix reference measurements were calculated to determine the dimensional changes. Data were subjected to ANOVA and Tukey's test (p<0.05). Results: For AB and BC distances, there was no statistically significant difference between CRIT and MRIT (p=0.0597 and p=0.2167, respectively). For AC, there was statistically significant difference between the techniques for the immediate storage time (p=0.006). In general, for CRIT the material showed expansion, while for MRIT it showed contraction. Conclusions: It was verified that the addition silicon impressions obtained by both impression techniques showed dimensional stability, except for the immediate time-point.