This paper presents a response evaluator which evaluates the response of the system and decides the control input to the system. It is applied to a base isolated structure with a semi-active oil damper whose damping coefficient has large or small values. Absolute acceleration of a top story and displacement of an isolated layer are the input signals of the response evaluator, and switching parameter for the damping coefficient is the output signal of the response evaluator. The response evaluator consists of a layered neural network, and a genetic algorithm is used to adjust the neural network parameters which decide the performance of the response evaluator. A computer simulation is carried out using several kinds of earthquake motions, and absolute acceleration and displacement of each story which are trade off relations of each other are reduced to the intermediate value of the response in case where constant damping coefficients are large or small. From the result of simulations, the effectiveness of the response evaluator is verified.