This study investigated the effect of modeling activities on scientific reasoning ability of elementary school students. Participants included 72 6th-graders from an elementary school in Kaohsiung who were nonrandomly assigned to an experimental group(n=36) and a control group(n=36). The scientific activity focused on chemical solutions that lasted for five weeks. After the activity, the two groups of students took various types of reasoning ability tests: (a) domain-specific (i.e., chemical solutions) reasoning test and (b) general science reasoning test (Lawson, 1978, 1995). Results of the study indicated that students in the experimental group scored significantly higher on both tests than students in the control group. Students in the experimental group scored higher on the domain-specific scientific reasoning test in general, and significantly higher on dimensions of volume conservation and control variables. Results suggest that when facilitating in-depth understanding of scientific phenomena, teachers may engage students in practices of imagistic reasoning and explanation to adjust the discrepancy between prediction and observation.