<p>Worker's presence sharing is important for smooth communication in a teleworking environment. The worker's state estimation system for presece sharing must be adapted to the environments of each worker. This research proposes a new method for creating a worker's state estimation system through cooperation between humans and machines. In many companies, workers telework from far away in various conditions, but are still required to effectively collaborate with one other. Many studies of human state estimation have reported a high estimation accuracy using machine learning. To create an estimator by the conventional method requires enormous amounts of learning data, and creating this data is a heavy manual task. The proposed method is create an estimator that reduces the manual task through the cooperation between people and computer systems. A prototype system was developed and tested in two workspaces. The proposed system reduced the manual task by approximately 60% and 80% in Workspace 1 and 2, respectively. The estimation accuracy was up to 90% in each workspace. The proposed method was effective in simplifying and reducing the manual task.</p>