The purpose of the study is to show the feasibility of a light-weight wearable electrocardiograph with three accelerometers for monitoring stress in everyday life. It consists of an amplifier, a band pass filter, a microcomputer with an AD converter, a triaxial accelerometer, and a memory card. An ECG and three accelerations are sampled at 1 kHz for more than 24 hours, maximum 27 hours with a default battery and a memory card. The heart rate variability analysis program detects R waves and determines R-R intervals for 24 hours by a data acquisition algorithm. An algorithm was also developed. It was designed to reduce motion artifacts induced by body movements on ECG. Autonomic nervous activity levels are calculated based on the heart rate variability spectrum analysis. The availability of the system was tested for three subjects for three days by replacing the battery and memory card every 24 hours under each environment. Both short time rhythms and a circadian rhythm of autonomic nervous system were clearly observed. The feasibility of the system for monitoring the long-term stress level is discussed. ©2010 IEEE.