Purpose / Context - It is important to estimate the number of people remaining in a room to allow for modification of the room environment and effective energy use. Methodology / Approach - A prototype of an automatic estimation system consisting of a general-purpose microcomputer and environmental sensors (e.g., a thermometer, hygrometer, and CO2 concentration sensor) was fabricated. We applied "Arduino 2009" as a root device and "XBee" as an extension. Results – According to experiments performed with this system, correct estimation was obtained once per eight tests. Key Findings / Implications – Although sufficient precision has not yet been demonstrated, our preliminary experiments led us to recognize that the system would be capable of error-free operation. We aim to achieve usable precision results by discarding the factors leading to incorrect estimations. Originality - This paper proposes a method to estimate the number of people remaining in a room of a house based on both CO2 concentration and humidity changes.