In this thesis, Digital Shoes system which senses the pressure of foot to infer human’s static and dynamic behavior was developed. In order to sense the human step sequences, the shoe hardware system was designed to measure the pressure on heel and toe, using two force sensing resistors on each shoe. Information obtained by sensors was transmitted via to desktop based classifier via Bluetooth. By means of the sensor data analysis, the human behavior was categorized into nine states by a rule-based classification. Applying pre-defined matching template, user’s states were categorized into: Walking forward, Walking backward, Running, Mark time, Standing, On knee, Toe stand, One leg stand, and On ground. The false state annunciation caused by weight shifting was reduced by considering the number of state transition in limited time. We examined the possibility of the least number of FSR for ambulatory monitoring. And it appears that the detection of human behavior using only two force sensing resistors is erroneous in some cases but possible. This system can be applied in the area of daily human behavior and sports activity monitoring and it also can be utilized as an input interface for entertainments.