This paper investigated the effect of role-playing on pupils' speaking skills in childhood education in Cross River State, Nigeria. The research design employed was a Pretest-Posttest control quasi-experimental design. The multi-stage random sampling technique was used to select 43 primary five pupils for the 2020–2021 sessions in two intact classes from two selected primary schools. The Speaking Ability Test (SAT), used as a Pretest (PRESAT) and Post-test (POSTSAT), was developed by the researcher and validated by experts. The reliability coefficient of the instrument obtained was determined using the Kuder-Richardson formula 20 (KR–20) after subjecting the tools to a trial test of 20 primary five pupils who are not part of the sample for the study but have relevant qualities as those in the study. The reliability coefficient was 0.80 obtained. Two hypotheses were formulated and tested at 0.5 level of significance. The Pretest was administered to subjects in both groups to determine their writing skills. The control group was taught speaking skills for three weeks using only the conventional method. The experimental group was also led to writing skills for three weeks with a role-play learning strategy. POSTSAT was administered to all the subjects at the end of the third week. The analytical results showed a substantial difference between students who learned speaking abilities through role-playing and those who knew them through traditional methods. The study also shows no significant difference between the performance scores of male and female pupils who were taught speaking skills using the role-play teaching method. It was therefore recommended, among others, that role-playing instructional practices be adopted in childhood education to improve the pupils' acquisition of speaking skills.