By means of contouring error control (CEC), the dynamics of robot is transformed into dynamic errors problem consensus with equivalent errors (EQs) in which such error dynamics is linearized around equilibrium point. However, it yields only locally asymptotical stability with invariant set of parameters in such dynamics. To reach consensus on asymptotically stability, globally, the feedback linearization associated with the set value of uncertainty is considered to enlarge an attractive region. The on-line self tuning proportional+derivative (PD) parameters by means of fuzzy membership function optimized by robust extended Kalman filter (REKF) is proposed to address the problem of the globally asymptotical stability of the dynamics of the robots under unpredictable environments. It could automatically determine on-line equilibrium point with such attractive region according to different processes without knowing a prior. The experiment results reveal that it yields the excellent performance under time-varying uncertainty. Moreover, theoretically, we explicitly derive globally asymptotical stability of the closed-loop system where attractive region is enlarged by REKF.