In teaching geometry, most instructors opt for direct demonstration with detailed explanations; however, under this kind of instruction students face considerable difficulties in the development of the reasoning skills required to deal with problems of a geometric nature. This study adopted a nonequivalent pretest-postest quasi-experimental design employing Polya's approach of four-stage problem solving using question prompts in conjunction with multimedia demonstration. Two classes of grade 7 students were randomly selected as the experimental group receiving instruction based on Polya questioning and two others were selected as the control group receiving instruction based on direct presentation. Our results revealed that the posttest performance in geometry reasoning of students receiving instruction based on Polya questioning was superior to that of students receiving direct presentation. In addition, students receiving instruction based on Polya questioning expressed a stronger sense of participation in the course than did students receiving direct presentation.