In order to stabilize the dust concentration for animal inhalation studies, two feedback systems were introduced. Dusts generated were passed through two cyclone separators and stored first in a storage chamber. Then they were aspirated into an ejector by compressed carrier air, mixed with filtered room air in a mixing box, and introduced into an exposure chamber. The first feedback system was for dust generation. Dusts were reaerosolized by on-off control using the output signals from a light-scattering dust detector connected to a storage chamber. The second system was for the control of compressed carrier air flow into an ejector. Compressed air flow was automatically regulated by a PID-controller in combination with a mass flow controller. The PID-controller processed continuous signals from another dust detector connected to an exposure chamber and fed control signals to a mass flow controller for regulation of the compressed air flow. This automatic control system could not only reduce hunchings of dust concentration following on-off control of a dust generator but also reduce overshooting of dust concentration in the exposure chamber following the detachment and dispersion of dusts adhering to the inner walls of the dust-supplying route. The system could also maintain a desired dust concentration in the exposure chamber for up to 180 days (20 hr/day, 7 days/week). In addition, this system could simultaneously introduce different doses of dust to other exposure chambers stably while maintaining similar size distribution of the dusts if additional identical exposure units were used. These results indicate that this dust inhalation system is suitable both for long-term studies and for investigating a dose-response relationship.