The utilization of dressed seed for spring sowing is a widespread practice to control some pests with reduced doses of chemical products. However some insecticides employed in maize seed dressing, namely belonging to the neonicotinoid family and fipronil, have been claimed to play a role in the decline of honeybees (Apis mellifera L.). Pneumatic drills used in maize sowing are charged with contributing to the dispersion of the abrasion dust produced by dressed seeds, favoring the contamination of the honeybee habitat. Different devices similar to air deflectors have been introduced on pneumatic drills in order to reduce dust drift. During previous field tests carried out by the authors during recent years reduction of dust concentration both in the air and at soil surface has been shown as a consequence of their application. As field tests are affected by the variability of environmental parameters (namely wind speed and direction) the results are not always reliable, comparable and of a general validity. This paper refers to a sowing simulation test system in which pneumatic drills can be tested at a fixed point under controlled conditions of the main environmental parameters. In the test area, protected by external influences, artificial wind conditions are created by means of a fan. The drill, suitably placed in the test area, operates the seed distribution "sur place" by means of an electric engine connected to the drill's driving wheel. A 22.5 m long sampling area, leeward with respect to the drill position, has been identified. Along the sampling area a series of Petri dishes has been placed, with the aim of capturing the depositing dust and providing the concentration of the active ingredients (a.i.) at ground level. At the same time, three air samplers with PTFE diskette filters have been used for the detection of the a.i. The test system has been used for the test of a pneumatic drill, equipped with and without air deflectors, using maize seed dressed with four a. i. (imidacloprid, clothianidin, thiametoxam, fipronil). The results showed regularly decreasing of the concentrations as distance increased, both in the air and at ground level. Moreover, the difference determined by the adoption of the drift reducing device (air deflectors) resulted clear and it can be quantified at around 50 % of the a.i. amounts observed without deflectors. Finally, the paper proposes a data processing method that, from the values observed at fixed point, provides the theoretical a.i. concentration behavior that would occur in field, under the same conditions of wind speed and direction and working speed. The obtained results are coherent with previous field test.