With the rapid development of robotic systems, an increasing number of various sensitive sensors are needed to meet the requirement of data collection for unmanned detection and monitoring technology. Among various external stimuli information, temperature sensing is an essential important function that can avoid high-temperature scalding injuries, achieve robot detection of environmental, and human-machine interfaces applications. However, most of the existing temperature sensors for robotic sensing are realized by rigid materials, which have the disadvantage of not being able to adapt to an arbitrarily curved surface. Also, most of the sensors are thermistor-based approaches, which faces challenges in developing a low-cost robotic system since external power is required to obtain the sensing signal. In this work, a self-powered and stretchable temperature sensor based on Triboelectric Nanogenerator (TENG) technology was developed. The principle of TENG is the coupling effect of triboelectric charge transfer and electrostatic induction effects, which can generate different magnitudes of electrical signals when subjected to external stimuli, including temperature. To this end, this study is a pilot effort toward validating the sensitivity of TENG electrical output on different temperature effects. The surface potential of materials is explored under different temperatures by Kelvin probe force microscopy (KPFM), which shows the feasibility of self-powered temperature sensor. The proposed temperature sensor is composed of highly resilient materials physiological saline as liquid electrode encapsulated with Ecoflex as triboelectric layer, which has the advantage of biocompatibility. The temperature sensing performance was successfully maintained at 200% stretch. Owing to its flexible, bendable and stretchable characteristics, the self-powered temperature-sensitive sensor has been successfully demonstrated on a robot hand, which can read the information and respond to different temperatures through the feedback system. Overall, this self-powered temperature-sensitive sensor has great potential to be an emerging tool for human-robot interaction and automatic detection.