To perform conversation or the like corresponding to the feeling of a conversation partner and to easily update the conversation or the like corresponding to a using purpose and conditions.

The conversation robot system includes a robot 1 provided with a voice input device 2, a voice output device 3, an expression output device 5 and an action output device 6 and a computer 4 provided separately from the robot 1 and wirelessly connected to it. The computer 4 respectively executes a step of analyzing voice inputted through the voice input device 2 and calculating a feeling parameter, a step of outputting prescribed voice through the voice output device 2 on the basis of a speech scenario, an expression scenario and an action scenario determined corresponding to the respective feeling parameters beforehand, a step of creating a prescribed expression through the expression output device 5, and a step of realizing a prescribed action through the action output device 6. The respective scenarios are stored in the memory 42 of the computer 4 so as to be updated.