Long-term experiments using organoids and tissues are crucial for drug development. Microfluidic devices have been regularly used in long-term experiments. However, microbubbles often form in these devices, and they may damage and starve cells. A method involving the application of negative pressure has been reported to remove microbubbles from microfluidic devices composed of polydimethylsiloxane; however, negative pressure affects the cells and tissues in microfluidic devices. In this study, a local microbubble removal method was developed using a microfluidic device with 0.5 mm thin polydimethylsiloxane sidewalls. The thin sidewalls counterbalanced the negative and atmospheric pressures, thereby localizing the negative pressure near the negatively pressurized chamber. Microbubbles were removed within 5 mm of the negatively pressurized chamber; however, those in an area 7 mm and more from the chamber were not removed. Using the local removal method, a long-term perfusion test was performed, and no contact was confirmed between the bubbles and the simulated tissue for 72 h.