Species biotic index (SBI), based on a fish assemblage, was applied to a stream assessment using long-term ecological fish data set (1996 to 2001) in Boryong dam area, which is located in the mainstream of Ungchun Stream of Chungnam province, Korea. According to the methods of Hilsenhoff (1988), the scores of tolerance guild assigned 10 classes to each species by its habitat and feeding guild, but modified current 7 criteria to 5 scoring standards due to unclear borderline among species. Relative abundance in the species number of upper stream guilds was only 7% of the total, whereas the abundance in the species number of middle to downstream upper stream guilds was 64%. Mean SBI, based on dataset in Site 1 during 1995-2001 averaged 5.10, which was judged as a "good" condition by the rank criteria of SBI. Before the dam construction, mean SBI in the Site 1 was 4.61, indicating a "good" condition, but after the dam construction, mean SBI was 5.60, indicating a "fair" condition. Trajectory analysis in the Site 1 showed significantly (One-way ANOVA, $F_{6,21}=3.26$, p=0.02) different among years, reflecting the changes of fish composition and population density by the dam construction, whereas Site 2 showed no significant changes ($F_{6,21}=1.00$, P =0.45) difference among years. Mean SBI prior to the dam construction in the Site 3 was 4.52 but after the construction, the value was 6.30, indicating a distinct difference between the pre- and post-dam construction. Trajectory analysis at the Site 3 supported this fact: Values of SBI showed significantly ($F_{6,21}$=14.37, p<0.01) different. Mean SBI was 4.67 in the Site 4, indicating a "good" condition in the health and the health rank was same as the sampling sites 1, 2, and 4. Trajectory in the Site 4 showed no significant ($F_{6,21}=2.35$, p=0.07) difference among the years. Overall, our trajectory analysis indicated that three of four sampling sites (sites 1, 3, 4) showed significant decreases (n=7, p<0.05) and that the proportions of sensitive species declined evidently in the sites 1 and 2 and the tolerant species increased in the dam sites. Our outcomes may be used as a key data for diagnosis of the long-term ecological impact in the future in the watershed.