Even though Translation Memories (TMs) are known to increase translation productivity, term consistency and quality, many scholars pinpoint that TM segmentation disrupts the Source Text (ST) linearity. Being an indispensable tool for translators and a novel tool in the Arab world, there is a growing need for an in-depth investigation of this feature. This research investigates the impact of TM segmentation on the Target Text (TT) in terms of cohesion and coherence. It is a mixed research design where participants translate two English user manuals into Arabic. The participants translated the first English smart phone user manuals into Arabic without the software program WordFast Anywhere (WFA) and translated the second through WFA. Three professional translation instructors at the AUC Language Department (AUC LD) assessed the language of the two texts produced in terms of cohesion and coherence based on the analytical rubric Farahzad (1992) and analyzed the protocols produced. The evaluation of pre-test, post-test and the Think-Aloud Protocols(TAP) revealed that TM segmentation negatively affects the cohesion and coherence of the TT