本研究旨在探討大學生身體意象與運動行為之長期關係，並探討維持運動行為之因素。以82學年度就讀於某國立大學學生研究母郡體，分層系統抽樣537份，採自填問卷收集兩次相關資料，初次施測在82年12月，兩週後(83年1月)進行再測，並於85年4月至5月間進行追蹤調查，追蹤時間為兩年半，共計回收有效問卷444份，以長期觀察追蹤。 所收集資料以頻率分布、平均值、標準差等描述性統計及、t檢定( t test)、配對t檢定(paired-t test)、相關分析(Pearson correlation)、共變數分析(analysis of covariance; ANCOVA)等統計方式進行分析。 研究結果發現;運動行為方面，僅只有30%的個案有達到每週3次，每次運動20分鐘的標準;男生在運動方面的頻率及時間均顯著高於女生，在追蹤測量時，有規律運動者不論男生、女生或整體樣本均有比率上升的現象。但考慮運動強度與時間之乘積後，發現運動行為量隨著年齡的增長而漸減。在排除性別因素後，體型認並非是影響運動行為的因素。就男女生分別分析，男性學生較滿意自我體型，而其運動行為不受到體型滿意度的影響。女生方面在追蹤測量的運動行為表現，明顯受到體型滿意度的影響，且為維持自己理想體而持續做運動。性別、身體意象與運動行為間有交互作用。 建議未來教育應針對性別及身體意象的差異，設計不同的運動方式，以增進適度的運動行為。藉由本研究的發現，作為日後研究身體意象影響運動行為的參考。

The main purpose of this study was to investigate the relationship between body images and exercise behavior and also to explore the factors that influence the maintenance of exercise behavior. The sample was taken from the undergraduate students of a university who enrolled at the academic year of 1993 by using the technique of stratified systematic sampling. The subjects were mailed a similar questionnaire at the year of the spring of 1996 for follow-up. Four hundred and forty four students who filled out the questionnaires of both surveys constituted the sample with a response rate of 74.5%. The data were analyzed by means of frequencies, means, standard deviation, t-test, paired t test, Pearson correlation and Analysis of Covariance. The results showed that: Only 30% of the subjects met he standard of exercising of three times a week, and 20 minutes each time; the male students were more active in terms of frequency and the time spent. The rat of the students who exercised regularly was higher at follow-up than the first-time measurement. Taking time and the strength of exercise behavior into consideration, the older the students grew, the less they exercised. It is true for both the male and the female. The perception of body shape was not the predictive factor of the students’ exercise behavior. Male students were more satisfied with their body image and the satisfaction did not have a lot of influence on their exercise behaviors. In the third year, female students who exercise for reasons specifically tied to the satisfaction with body shape were primarily to obtain an ideal body for developing exercise patterns. The interaction was among gender, body image and exercise behavior. The authors suggested a need for gender-specific tailored of message on body image and the type of exercise, also to improve adequate amount of the importance of exercise. Based upon these results further research of the mechanisms by which body image influences exercise behaviors is warranted. The authors suggested the need to design the various kinds of exercise based on sex and body image, so to improve the condition of doing exercise.