Background Current soft endoscope intervention mainly depends on the skills of endoscopists, especially for the orientation of the endoscope within the human tortuous lumen. This dependence causes inefficient and time-consuming surgery. Methods A human-robot cooperation navigation strategy is proposed to perform the intervention of soft endoscope. The navigation strategy firstly detects the cavity center by processing the image of endoscope, then an auto-orienting strategy is presented to automatically adjust the endoscope pointing to the detected cavity center. Results Human phantom experiments are carried out to validate the effectiveness and feasibility of the navigation strategy. The results show that the cavity center is accurately extracted with whether the cavity is occluded or not and the auto-orienting strategy could well point the endoscope around the cavity center. Conclusion The proposed human-robot cooperation navigation strategy releases endoscopists from the time-consuming operations and improves the safety and automation of intervention procedures.