The Boost/Buck converter was introduced into the energy-harvesting circuit to coordinate dynamic performance and energy-harvesting performance of energyharvesting suspension system. The non-linear mathematical model of DC—DC converter was described by the summation of steady-state components and disturbance components. Considering the mapping capability of neural network and anti-interference of FIR filter，adaptive off-line neural network inverse controller was designed. The controller was composed of off-line neural network inverse controller and adaptive controller. The off-line neural network inverse controller was used to provide steady control input with steady components of DC—DC converter as training samples, while the adaptive controller was used to provide compensation dynamic control generated by parameter variations and uncertain interference. On this basis, the converter sample was made and the rapid control prototyping experiment was performed to validate the performance of the controller. To meet the suspension vibration isolation performance and harvest vibration energy to a large degree, the suspension control parameters was decided. And a linear motor was tested to obtain its characteristic parameters. Simulations were applied to a two-degree of freedom energy-harvesting suspension. The results showed that the linear motor can produce the desired electromagnetic damping force with the proposed controller and improve vehicle comfort. It also indicated that self-powered control with energy harvested from suspension was attainable.