This study was motivated by the low mathematical problem-solving ability of fourthgrade elementary students in fraction material, particularly in conceptualunderstanding and systematic completion of solution steps. It aims to describe theimplementation of the Peer teaching method assisted by manipulative media and toexamine its effect on students' mathematical problem-solving ability. Using aquantitative quasi-experimental approach with a nonequivalent control groupdesign, the study involved 60 fourth-grade students of SDN Tegalpanjang, dividedequally into an experimental class and a control class. Data were collected throughpretest-posttest written tests and learning-implementation observation sheets, thenanalyzed using normality, homogeneity, and independent samples t-testprocedures. The results showed that the implementation of learning in theexperimental class improved consistently across meetings and reached a highoverall category (71.67%). Welch's t-test on the posttest scores yielded asignificance value of 0.000 (< 0.05), indicating a significant effect of the Peerteaching method assisted by manipulative media on students' mathematicalproblem-solving ability. The greatest improvement occurred in the problemunderstanding indicator, while the indicators of implementing the solution andreviewing the results still require further reinforcement. Overall, this method provesto be an effective alternative for improving elementary students' mathematicalproblem-solving ability.