This study aimed to investigate the efficiency of a training program meant to improve the working memory of learning disabilities female students at Muscat, Sultanate of Oman. The sample consisted of (10) female 5th graders with learning disabilities from Bowsher and Seeb. The sample was divided into experimental group and control group with five students in each. In this study, I used Raven's Progressive Matrices test to diagnose learning disabilities students and verify parity of both experimental and control groups. I also used Children WM Functions Battery prepared by (Alloway, 2007), translated by (Suleiman, 2010) in order to measure parity between both groups as well as the efficiency of the training program. The program was applied over learning disabilities female students after being authenticated by a group of arbitrators. The program which lasted for (2) month included (16) training sessions by rate of (2) sessions weekly. The battery reliability was verified via (Al-Fori, 2015) study through test-retest method and internal consistency (Cronbach's Alpha) while study validity was verified through (Suleiman, 2010) study using Criterion-related Validity and Concurrent Validity. Mann Whitney test was used for the purpose of testing study hypotheses. The results concluded that there are differences between scores of control group and scores of experimental group on the post measurement. They were in favour of the experimental group. Moreover, experimental group members were subject to Wilcoxon test for prior and post measurements of working memory. The results demonstrated that there are differences between experimental group scores in the prior measurement and post measurement and they were in favour of the post measurement. However, there were no differences between the group scores in post measurement and postponed measurement.