Abstract: A type of ventilation method, multi-module adaptive ventilation, hereinafter referred to as MAV, is proposed. At the same time, it studies to control the diffusion of pollutants in classroom buildings when facing with changes in indoor scenes and changes in the location of pollution sources and the ability to remove contaminants. Computational fluid dynamics technology was used to simulate the distribution of pollutants in the classroom in normal class and group discussion scenarios, and to change the location of pollution sources in the normal class scenario. The tracer gas (CO₂) was used to simulate the pollutants produced by the cough of an infected person. Five different MAV modes and conventional mixed ventilation (MV) modes were selected to compare their pollutant concentrations and removal rates in the respiratory zone of personnel in the face of different pollutant source locations and different scenarios in the classroom. Results show that the relative concentration of pollutants in the breathing zone is lower and the exclusion rate is higher with the appropriate MAV mode, in which the pollutant removal rate (CRE) in the breathing zone of the MAV parallel mode is 7.52 times that of the MV mode in the group discussion scenario and the location of the pollutant source is P3, and the relative pollutant concentration C_r in the breathing zone is 52.4% of that in the MV mode, indicating that the appropriate MAV mode can better adapt to the scenario and the change of pollutant source location.