<![CDATA[ Arabic language learning in Islamic boarding schools (pesantren) plays a central role in students' religious and academic development. However, passive learning environments and the dominance of conventional methods continue to hinder optimal learning outcomes. This study examines the effect of the Think Pair Share (TPS) cooperative learning model on Arabic language learning outcomes among eighth-grade students at MTs Raudhatussalam Rokan Hulu. A quasi-experimental design with pre-test and post-test control group was employed. Subjects consisted of 38 students: 23 in the experimental class and 15 in the control class. Data were collected through observation sheets and learning outcome tests, then analyzed using IBM SPSS version 27 through descriptive statistics, Shapiro-Wilk normality tests, Levene homogeneity tests, and independent sample t-tests. Results showed that the experimental class mean increased from 63.91 to 78.70, while the control class mean increased from 70.00 to 79.33. The independent sample t-test yielded t = -12.493 with a significance value of 0.000 < 0.05, confirming that Ha is accepted and Ho is rejected. Observation records across six sessions showed consistent implementation improvement from 78% to 94%, with an overall average of 87%. These findings demonstrate that the TPS model significantly improves Arabic language learning outcomes by actively engaging students through structured individual thinking, paired discussion, and whole-class sharing, making it a viable and recommended instructional strategy for Arabic language teachers in pesantren settings. Keywords: Think Pair Share, cooperative learning, Arabic language learning outcomes, pesantren]]>
