The purpose of this research was investigated whether the different of problem-posing types can improve elementary school students’ ability. We also conducted a posttest to determine whether the different of problem-posing types would have positive effect on fifth graders’ learning affect. A quasi experimental non-control group designed was used in this design. Participants were 96 numbers of fifth-grade of elementary students in Kaohsiung. There were three classes assigned to experimental groups and control group. The experimental groups received structure materials and semi-structure materials while the control group received free problem-posing materials. After four times practice, we conducted posttest. In quantitative data analysis, the results show that, the experimental group, who received structure materials, demonstrated better problem-posing ability than the control group. In posttest, the experimental group, who received structure materials, rendered better performance than the control group in learning affect.