This article investigates the event and self-triggered adaptive output feedback control of a manipulator suffering from limited knowledge of states and dynamics, to realize the trajectory tracking with less communication occupation. In this control scheme, the configuration of co-located observer and controller with discontinued output feedback is considered. To guarantee the convergence of observation and control errors with few events as much as possible, an adaptive event-triggered mechanism based on model estimation is constructed to compensate for the error accumulation produced by the intermittent open-loop. Based on the model state, adaptive backstepping method with network estimation is used for deriving the controller, to solve the control stability under uncertainty of system dynamics. Aiming at removing the “derivative explosion and singularity” of discontinuous virtual signal, a first-order filter is incorporated to get the smooth approximation of the virtual signal, and an additional self-adaption signal is designed for the filtering error compensation. In view of the state updating at event instants, a gradual updating method is designed such that the state jumping-induced chattering instability could be handled. With the above designed method, a dead-zone event-triggered condition with the relative threshold and variable tolerance boundary is built to avoid Zeno-behavior. Furthermore, an easy-implemented self-triggered mechanism is also constructed. Finally, the Lyapunov function is utilized to derive the setting principle for the stability of the system, and the simulation is given to show the validity of the proposed control method.