The objective of this study was to evaluate the effect of the contamination of the dentin tissue and the methods normally employed in the cleaning of the substrate on the bond strength obtained at the resin modified glass ionomer cement (RMGIC) / dentin interface. For this purpose, 168 bovine teeth were selected, which were worn to dentin tissue, and were then divided into twelve groups (n = 14), according to the contaminating agent and the cleaning technique NN - without contamination (N) and without substrate cleaning (N); NW - no contamination (N) and spray / water cleaning (W); NP - without contamination (N) and cleaning with 26% polyacrylic acid (P); SaN - contamination by saliva (Sa) and without substrate cleaning (N); SaW - saliva contamination (Sa) and spray / water substrate cleaning (W); SaP - saliva contamination (Sa) and substrate cleaning with 26% polyacrylic acid (P); BN - blood contamination (B) and no substrate cleaning (N); BW - blood contamination (B) and cleaning the substrate with spray / water (W); BP - blood contamination (B) and substrate cleaning with 26% polyacrylic acid (P); HN - contamination by hemostatic agent (H) and without cleaning of the substrate (N); HW - hemostatic agent contamination (H) and spray / water substrate cleaning (W); HP - hemostatic agent contamination (H) and substrate cleaning with 26% polyacrylic acid (P). All specimens were restored with the Riva Light Cure (SDI) encapsulated ionomeric material and, afterwards, sticks of approximately 1.0 x 1.0mm were obtained, submitted to the microtraining test before and after performing 10000 thermal cycling (TC). The fragments were analyzed for the fracture pattern and the representative specimens of each group were taken by scanning electron microscopy at a magnification of 1000x. Values were submitted to two-way ANOVA and Tukey post-test for parametric data and Kruskal Wallis and Holm-Sidak post-test for non-parametric (α = 0.05). In the immediate analysis all the groups contaminated without cleaning had lower values of adhesion, when compared to the control. In the saliva contamination, the group that did not receive any cleaning and the group washed with air / water spray had the lowest values of bond strength, however the refurbished group presented similar values to the control. Analyzing the contamination by blood in the immediate time, the proposed treatments were not as effective in cleaning the substrate. Already in the groups contaminated by hemostatic the reconditioning was similar to the control group. After TC, in the contaminated groups without cleaning, only the hemostatic was similar to the control. The groups contaminated by saliva, however, maintained the same pattern of immediate analysis. And those contaminated by blood that received cleaning were similar to the control, which presented significant reduction in its value differently from the immediate time. Finally in hemostatic contamination all groups were similar to each other. It was concluded that the contaminating agents tested impaired the immediate adhesion of the ionomeric material to the dentin substrate. Substrate reconditioning was effective for saliva and hemostatic contaminations and the bond strength of the material to the dentin was reduced after aging(AU)