In this study, inclined broken roughness is experimentally investigated to enhance the thermohydraulic performance of solar air heaters. Current article focuses on a fuzzy-based structure as a substitution method for predicting the thermohydraulic performance. A separated fuzzy inference system for each smooth and roughened plate along with an integrated one designed for both plates was selected as the strategy of modeling. In addition, utilizing temperature and velocity features with their generality in all solar air heaters were suggested as a cutting-edge solution to dominate complexity restrictions of geometrical roughness parameters. Throughout the experiments, triangular membership functions obtained better agreement with experimental data than Gaussian functions except for the friction factor of the roughened plate in the separated method. Sugeno structure demonstrated better forecasting ability than Mamdani. Additionally, Nusselt number showed better applicability in being predicted more easily by the considered fuzzy structures rather than the friction factor. Moreover, the system constructed based on the Gaussian membership function showed higher accuracy in forecasting the roughened plate parameters. The least mean square error of the separated method for the Nusselt number and the friction factor of the smooth plate were 2.5477 x 10(-04) and 8.1115 x 10(-04), respectively. Furthermore, these values were equal to 2.0218 x 10(-04) and 7.5150 x 10(-04), for the Nusselt number and friction factor of the roughened plate, respectively. Having considered the thermohydraulic performance, the least mean square error of the separated method was equal to 8.6255 x 10-04. The obtained results approved that the fuzzy method is a significantly efficient method for anticipating the parameters of solar air heaters.