<![CDATA[Speaking fluency become a major obstacle for students in English as a foreign language (EFL) classrooms due to limited opportunities to communicate spontaneously. This research aimed to find out the effect of randomized topic-based speaking tasks to improve students’ speaking fluency. The study employed a quasi-experimental pretest-posttest non-equivalent control group design involving 62 eighth-grade students aged 13-14 years old from a middle school in Malang. An experimental group participated in randomized topic-based speaking tasks activities using flipped bottle media, and the control group learned with the usual textbook method. Data came from speaking tests and assessed using a fluency scoring rubric followed by Brown’s (2004) criteria for speaking. The findings showed that the experimental class obtained a better post-test mean score (M= 58.83) compared the control class (M= 43.13). Furthermore, the Mann-Whitney U test make a significant difference for both groups (p= .001), indicating that the treatment positively affected students’ speaking performance. These findings suggest that flipped bottle media can serve as an effective pedagogical tool by providing randomized speaking topics that encourage spontaneous language production and meaningful classroom interaction. Therefore, integrating simple physical game media into task-based speaking activities may offer a practical alternative for improving speaking fluency, particularly in EFL classrooms with limited technological resources.]]>
