This study was designed to investigate the possibility that there might be a significant improvement on the students' reading comprehension achievement following the treatment through Cohesive Device Recognition Skill-Building Exercises. It was conducted under the principles of the one group pretest-posttest (repeated-measures) design. 4 J students of Economics Education of FKIP-Unila attending English as a general subject were chosen as the subjects. These students were introduced and trained with the skill and encouraged to firstly recognize the cohesive devices employed in the texts, then classify their types and functions across the sentences and paragraphs, and finally to check their comprehension about the texts on hand. As the design suggests, this research administered pretest and posttest as the main instruments in collecting the data. The data were analyzed using Repeated Measures-Matched T-test. The result shows that t-observe (6./79) is higher than Hable (2.704) which means that there is a significant improvement on the students' reading comprehension achievement following the treatment. This seems to suggest that this technique worked well and was found effective in this stud)~ Since the study was conducted to an intact class and even though the test allows us to generalize, the design of the study precludes such generalization. Therefore, further study involving randomly selected and randomly assigned subjects would certainly provide more interesting findings.