In this paper semi-active control of the suspension of an all-terrain vehicle is considered. A seven degree of freedom suspension model is presented first. A fuzzy approach for controller synthesis is then proposed. Expert knowledge is stored in the form of IF-THEN rules. The Takagi-Sugeno inference system is employed, with triangle membership functions. The fuzzy system output is the damper coefficient. In contrary to many other control algorithms, the presented fuzzy algorithm does not require inverse modelling of the magnetorheological damper. Instead, some scaling parameters are set. They can be chosen experimentally, or a bio-inspired strategy can be applied. The fuzzy control is then compared with the Skyhook control in simulations, in terms of road holding and driving comfort indicators. Obtained results are similar. However, lack of necessity to use an MR inverse model allows the fuzzy system to provide successful performance in case of different operating conditions, what is an important benefit.