Objective: This study evaluated the biocompatibility of two dentin bonding agents (Clearfil Liner Bond 2 and Scotchbond Multi-Purpose) applied in human dental pulps and cell cultures. Method and materials: In vivo: Twenty human third molars that were scheduled for extraction were used. After cavity preparation, pulp exposure was achieved with a carbide bur. Hemorrhage control was obtained with saline solution. In 16 teeth, adhesive pulp capping was performed and the cavities were sealed with resin composite. In the control group (n = 4), pulps were capped with Ca(OH)2 and the cavities were sealed with IRM. Teeth were extracted 30 or 90 days following treatment and prepared for histological examination and bacterial detection. In vitro: materials were applied in Petri dishes, where NIH-3T3 cells were plated. The cells were counted 2, 4, and 6 days after plating to obtain the growth curves and to determine cell viability. All data were submitted to statistical analysis. Results: In vivo: Dentin bridge formation was seen in all teeth capped with Ca(OH)2, without an inflammatory response. Mild inflammatory responses and dentin bridge formation after 90 days were observed in 50% of specimens treated with Liner Bond 2. Pulps treated with Scotchbond Multi-Purpose presented mild to severe inflammatory response, and no mineralized tissue formation was detected. Bacteria were not disclosed in any specimen. In vitro: The cytotoxicity was similar between the two bonding agents, and both had statistically higher cytotoxic effects (P 0.002) than Ca(OH)2. Conclusion: Ca(OH)2 produced pulp healing in all teeth and exhibited lower cytotoxic effects than both adhesive systems; however, pulp healing was also observed under Liner Bond 2.