Complex mechanical systems, such as tower cranes, are known to be highly nonlinear, under-actuated, and non-colocated, which makes their closed-loop control very challenging. The interconnected components of these systems undergo complex dynamic phenomena, such as friction, which lead to energy or momentum transmission delays. The complexity of such systems is further complicated by external disturbances and nonlinearities resulting from using hydraulic and/or electrical actuators, mechanical joints, gears, etc., which result in the formation of dead-zones, backlash, and hysteresis. A dead-zone, which constitutes a significant non-smooth nonlinearity, severely limits the performance of many mechanical systems such as the tower crane. Previous works on the control of tower cranes were based on accurate determination of their actuated states. In this work, a robust control technique based on adaptive fuzzy theory is investigated for anti-swing and trajectory tracking of tower crane systems. The system is subject to uncertainties in parameter parameters, time delays, external disturbances, and unknown actuator nonlinearities. The unknown actuator nonlinearities, from the jib and tower motors, are characterized by dead-zone bands (as opposed to the typical crisp dead-zone functions). First, fuzzy logic systems with on-line adaptations are utilized to evaluate the unknown nonlinear functions. The proposed control scheme uses the H control technique to develop compensators to overcome the effects of parameter variations, time delays, external disturbances, and unknown actuator dead-zone band nonlinearities. The proposed control scheme ensures the stability of the closed-loop system and achieves desired tracking precision such that the states of the tower crane system are ultimately uniformly bounded (UUB) and guarantees an H norm bound constraint on disturbance attenuation for all admissible uncertainties based on the Lyapunov criterion. Simulation results show the validity of this approach for the tower crane system.