This research was the first research integrating decision making trial and evaluation laboratory (DEMATEL) and structural equation modeling (SEM) to analysis factors, including interest in science, value in science and confidence in science, affecting female students' science achievements from the Trends in International Mathematics and Science Study (TIMSS) 2011. The research constructed two causal models base on mutual effect of value in science and confidence in science by DEMATEL analysis. The causality of two factors in model 1 was value in science affected confidence in science and model 2 vice versa. According to the results of SEM analysis, fit indices of CFI, GFI and AGFI in model 1 were 0.851, 0.796 and 0.745; in model 2 were 0.818, 0.772, and 0.873. And results showed that female students had better fitness in model 1; in other words, the causality of value in science to confidence in science might have an important effect on the science achievement of female students. Lastly, these findings could provide not only a different perspective for future researches but also a useful insight as a reference for policy making and improvement by relevant decision makers.