This study aims to evaluate the effects of a non-routine problem solving instruction on the brain hemisphericity of the 7th graders (aged 13-14 years). One sample experimental design was used and twenty-one seventh graders participated in the study. Validated by Kök (2005), Brain Hemisphericity Inventory (BHI) was conducted to the experimental group as a pre-test. Participants were given an instruction which lasted for 18 lesson hours, and participants solved 80 non-routine problems throughout the instruction. The BHI was then applied to the experimental group as a post-test. The findings showed that there were significant differences between pre- and post-test scores of the participants. Although participants were mostly left-brained before the instruction, the number of left-brained participants decreased, and the number of right-brained students increased after the instruction. Findings indicated that non-routine problem solving may have a potential to change brain hemisphericity of the students. Based on that, it was concluded that amount of time and place given for non-routine problems in text-books and learning environments should be increased.