The super-aged society is becoming a reality and maximizing the ability of the elderly to live independently and reducing the burden on caregivers have become urgent issues. In this study, we propose to establish an emotion-reading nursing care environment based on facial expression recognition to reduce the long-term work pressure of caregivers and achieve the goal of improving the quality of life of the elderly. The system automatically adjusts the environment through facial expression recognition to guide the emotions of the care recipient. It uses the expression images captured by a computer and a camera lens, and uses the Python-trained program for emotion recognition. The captured identification signal is transmitted to a LabVIEW platform via Transmission Control Protocol/Internet Protocol (TCP/IP). The LabVIEW control program can guide emotions and corresponding environmental conditions, including color, sound, smell, and temperature, in accordance with the different emotional events received, and output the processing results to the real-time embedded evaluation board MyRIO to link devices such as a liquid crystal display, a servomotor, and a full-color LED. In this study, we demonstrate the feasibility of an adaptive and optimized nursing care environment based on facial expression recognition. However, through experiment, we also understand the constraints of the system and differences from the theory.