Students with learning disabilities have difficulty in science literacy because of the extensive amount of abstract concepts in comprehending science content. An adapted alternating treatments design with an extended baseline was used to compare the effectiveness and efficiency of graphic organizers (GOs) enhanced with predicting-observing-explaining (POE) (Module-I) and GOs without POE (Module-II) on science achievement of six middle school students with learning disabilities. Results indicated that both procedures were effective to improve performance on all dependent variables for all participants. However, the result of efficiency data showed that Module-I seemed to be more efficient across four students in terms of number of intervention session through criterion, number of intervention session with 100% accuracy performance, and number of correct responses in maintenance sessions. The social validity findings were positive overall. On the basis of an evaluation of findings, implications and future research needs are discussed.