本研究旨在探討旗山老街遊客的族群結構、遊憩特性、並建構及驗證老街遊客之遊憩行爲模式。採自編的問卷進行研究，內容包括旅遊意象、行爲態度、遊憩滿意度、行爲意圖、旅遊特性及遊客基本資料等六大項。從2005年2月26日、27日、28日及3月5、6日，以至旗山老街之遊客爲研究對象，採用便利抽樣法，實施問卷調查。共發放478份問卷，扣除廢卷28份，實得有效問卷450份，抽樣誤差爲4.6％（95％信賴水準下）。將所收集的資料以SPSS 12.0的敘述統計程序，探討解統計屬性與旅遊特性等分布情形，利用LISREL 8.52進行驗證性因素分析及結構方程式的實證研究，以最大概似估計法(Maximum Likelihood Estimation)對本研究模式中各組變項間之因果關係進行估計。結果顯示受訪遊客以男性居多（57.8％），年齡以21－29歲居多（50.4％），婚姻以未婚者居多（76.0％），學歷以大專（學）者居多（70.4％）)，職業以學生（41.8％）)居多。受訪遊客之旅遊屬性以偏向懷舊之旅最高（45.1％），同行伴侶以朋友、同事、同學者居多（55.6％），旅遊次數以三次造訪者居多（55.8％）。結構方程模式之實證研究得知遊客的旅遊意象及行爲態度顯著的直接影響遊憩滿意度，間接影響遊客的行爲意圖；遊客的滿意度顯著的直接影響遊客的行爲意圖。

This study explored tourists' structure, tourism characteristic, and constructed the behavioral model using latent variables of tourism image, attitude, satisfaction, and behavior intention of Chishan old streets tourists. A total of 478 questionnaires were administered from February to March 2005, 28 questionnaires were invalid owing to being incomplete, leaving a total of 450 valid questionnaires. The descriptive statistics of socio-demographic was analyzed by SPSS 12.0. The confirmatory factor analysis and structural equation model analysis were analyzed by LISREL 8.52 for windows. All the parameters were estimated by maximum likelihood method. Mainly descriptive findings were as follows: respondents were 57.8% men; 70.4% aged from 21 to 30 years old; 76% unmarried. Empirical results indicated that both the tourism image and attitude affected the satisfaction intention directly and affected the behavior intention indirectly. Tourists' satisfaction affected the behavior intention.