The purpose of this study was to investigate the relationships between Sensation-Seeking, Creativity, and Delinquency, and to test Farley's theory. It was predicted that creativity and delinquency would be shown to be functions of physiological arousal and the stimulation-seeking motivation. These hypotheses were tested on two groups. The sample consisted of 117 male Chinese juvenile delinquents and 398 male non-delinquent Chinese adolescents. Five measures were used: Chinese versions of Zuckerman's Sensation Seeking Scale (SSS-23), Environmental Approval Index, Active Check List (ACL), Torrance's Creativity Test (TTCT), and Khatena's Something About Myself (SAM). Sensation seeking was measured by the SSS-23, containing three factors: Thrill & Adventure Seeking (TAS), Socially Disapproved Behavior (SDB), and Personal Relationship & New Life Style Seeking (PNS). Environmental approval consisted of family, peer, and school supports. Delinquency was measured by Delinquency Scale (DS), one of the ACL subscales. Torrance's Unusual Uses and Parallel Line activities, and SAM served as measures of creativity. SAM consisted of five subscales: Problem Solving Behavior (PSB), Open to Experience(OE), Artistic Performance (AP), Self-Preserving Behavior (SPB), and Self-Confidence (SC). The results of the study indicated that: (a) Delinquents' scores were higher than the non-delinquent group on DS and SAM, but lower on TTCT. (b) High sensation-seeking adolescents' scored significantly higher than low sensation-seeking adolescents on DS, SAM, and TTCT. (c) There was a significant difference between high, medium, and low environmental approval groups on DS. The medium group's score was higher than the others. For creativity, only Problem Solving Behaviors in SAM subscales indicated a significant difference between high and low environmental approval groups. (d) The results of Stepwise Multiple Rregression showed that both sensation seeking and environmental approval could predict creativity as well as delinquency. For the delinquents' sample, it could predict 36.7% delinquency, and between 3.7% and 22.4% creativity. For the non-delinquent sample, it could predict 36.2% delinquency, and between 1.4% and 13.5% creativity. (e) Although creativity and delinquency have common origins in hyperarousal, the type of sensation-seeking motivation was different. In SSS-23 subscales, if higher TAS, SDB, and PNS would enhance creativity, higher SDB only would increase delinquent behavior. In the conclusion, it is suggested that delinquency needs a new definition in modern society. Futher research in the area is also suggested.