In this paper, the issue of actuator-structure interaction in dynamic testing of structures is considered. The problem is approached from the novel standpoint of impedance control. It is shown that an effective strategy to design controls for dynamic testing is by designing the test system impedance. It is also shown that this can be achieved using feedforward compensation. The analysis is carried out in the context of displacement controlled dynamic testing, when the tested structure has a high and nonlinear stiffness. It is demonstrated that stable and accurate dynamic testing can be achieved using the proposed strategy, when this is not possible using traditional feedback control techniques. Furthermore, the impedance control and feedforward strategies are applied in the context of hybrid simulation, a technique of coupling computational and physical substructures applied in earthquake engineering. Here, a delay compensation scheme is necessary in addition to feedforward. Experimental results are presented that demonstrate both improved dynamic testing performance when impedance control is employed, and its applicability in hybrid simulation. In this paper, the issue of actuator-structure interaction in dynamic testing of structures is considered. The problem is approached from the novel standpoint of impedance control. It is shown that an effective strategy to design controls for dynamic testing is by designing the test system The analysis is carried out in the context of displacement of controlled dynamic testing, when the tested structure has a high and nonlinear stiffness. It is demonstrated that that stable and accurate dynamic testing can be achieved using the proposed strategy, when this is not possible using traditional feedback control techniques. when this impedance is not possible using traditional feedback control techniques. a delay compensation scheme is necessary in addition to feedforward al results are presented that demonstrate both improved dynamic testing performance when impedance control is employed, and its applicability in hybrid simulation.