This study intended to measure the effect of Task-Based Learning (TBL) on lower secondary school students' understanding of chemical reactions. The study employed a quasi-experimental pre and post-test research design with eight intact classes of 369 students purposively selected from eight schools in two districts in Rwanda. The experimental group of four schools was exposed to the TBL method, while a control group of the other four schools was exposed to the conventional teaching method (teacher-centered). A chemistry test of a Pearson product-moment reliability coefficient of 0.643 was developed and used in both groups. The results of repeated-measures ANOVA revealed a significant effect of treatment on students' understanding of chemical reactions (p < 0.001) with a medium effect size (d = 0.357) in favor of the experimental group. Gender and school location variables were also analyzed. It was found that the interaction effect of experimental and gender was not significant (p > 0.05, d = 0.010). However, the effect was significant with the school location (p < 0.05, d = 0.026) in favor of students studying in rural schools. The results imply that TBL method improves students' understanding of chemical reactions. Recommendations were given to educational stakeholders to train teachers in this method and teachers to use it in chemistry teaching.