One of the cleaning methods for agricultural materials is based on aerodynamic properties. Pneumatic cleaners are developed on this method. The purpose of this study is to predict the parameters such as fan angle, air velocity, and tunnel length, which are used in the design of pneumatic cleaners, through the multivariate adaptive regression splines (MARS) method. Some parameters have been estimated using the MARS method in order to use pneumatic cleaners under optimum conditions and adapt them to automation systems. The cleaners have a collection box which was installed at the outlet of the storage. Two different product collection boxes of 400 mm (defined as the first box) and 800 mm (defined as the second box) from the storage outlet section were used. From the results obtained, it was observed that the first box R-2 was higher. When looking at the cross validation, it was observed that the results of the first box were more acceptable. With this study, MARS equations were used to obtain dependent variables at desired values. Using these equations, independent variables have been demonstrated to be identifiable. In the application results obtained, cleaning efficiency values were obtained in a wide range. While cleaning efficiency values reached up to 100%, the loss rate was found to be very high. Independent variables have been made identifiable to reduce the loss rate. The highest and feasible of these values were determined by MARS as 41 degrees fan angle and 15 m/s air velocity in order to be able to apply at 97% CE and 1% LR determined for the first box. The MARS method allows for the use of more dependent and independent variables. Usable equations were obtained as a result of statistical analysis. More precise values can be obtained with these equations. It will contribute to the design of the parameters of the machine manufactured, such as speed, angle, and feeding amount.