In visual servo system, there are problems on time delay caused by image processing and difference between image processing period and control period. In order to overcome these problems, an estimation method of image features has been applied to a fixed camera, which realizes visual servo with control period. Since the compensation of delay utilizes image Jacobian matrices, it is affected by calibration error. Therefore, this paper proposes a new visual servo system for manipulators and active cameras using the estimated image features with a simple online calibration. In order to reduce estimation error of image feature, the proposed method is calibrating each element of image Jacobian matrices from error of the estimated image feature. It is simple algorithm and few calculation amount compared to calibrating each parameters correctly such as internal camera parameters, orientation of manipulator and camera, and so on. The effectiveness of the proposed strategy is confirmed by tracking simulation of a moving object by a manipulator and an active camera