Speaking is one of the speaking skills that students want to achieve in learning, one of which is Arabic subjects at MI Tanwirul Hija Selorejo Baureno Bojonegoro, especially in class VI, which is still low on Arabic speaking skills. Because in learning Arabic in class VI still not using learning strategies or learning models that are interesting and varied. One of the efforts offered as an alternative in solving the above problems is the application of a cooperative learning time token learning strategy to support students in the process of learning Arabic speaking skills. This type of research uses a quantitative approach with experimental research methods (pretest post test one group design) by comparing the scores of students before being given treatment (Treatment) with after being given treatment. The results of this study show that the t_count is 66.708 and db is N-1 = 27-1=26, while the criteria for acceptance or rejection of H_o are H_a accepted if t_test> t_table at a significance level of 0.05 (5%) and 0.01 (1 %), then use the provisions of the distribution table "t". After being consulted on the T distribution table, precisely in the 5% significance level column, t_table = 2,056 and 1% significance level obtained t_table = 2,779 using a two-way hypothesis test form, then the conclusion is the analysis result of t_count is greater than t_table, i.e. 66.708 > 2,056 and 2,779. It is concluded that H_a is accepted and H_o is rejected. Thus, it can be concluded that there is a significant effect of the time token learning strategy on improving the Arabic speaking skills of the sixth grade students of MI Tanwirul Hija Selorejo Baureno Bojonegoro for the 2021/2022 academic year.