The tracking control accuracy of the piezoelectric actuator (PEA) is limited due to their inherent hysteretic nonlinearity. Direct drive of PEA on a positioning stage with friction force causes control problems in static errors, varying dynamic frictional force, limit cycles and stick-slip, et al. An approximated PEA model is synthesized based on linear transfer function with two uncertainty parameters for time delay and frictional force effect. The frictional model of the motion stage in the presliding and sliding regimes is considered thoroughly. The H-infinite tracking controller is designed for compensating the hysteresis delay and frictional force in PEA actuated stage during positioning. The Iterative Learning Control (ILC) is implemented to reduce the unmodelled repetitive error from the frictional characteristics. Numerical simulations and experimental tests consolidate that the RMS positioning error can be close to the hardware reproducibility and accuracy level. Experimental results show the controlled piezo-stage can be potentially used for nano technology applications for precision engineering in industrial systems.