This study applied the Technology Acceptance Modes (TAM)framework to investigates university EFL students' acceptance of DeepL machine translation tool, focusing on how frequency of use influences their perceptions. A descriptive qualitative approach was used involving purposive sampling of four students in semi-structured interviews. A short questionnaire was administered beforehand to classify participants into two categories, two frequent and two infrequent DeepL users. of the machine translation tool DeepL in the context of learning English as Foreign Language (EFL), involving four participants (two frequent users and two infrequent users). The results showed that frequent users found DeepL valuable for vocabulary acquisition, translation accuracy, and academic writing support, whereas infrequent users highlighted limitations such as the lack of a paraphrasing feature and issues with formality. These differences suggest that usage frequency significantly impacts perceived usefulness and ease of use of the tool. This study is among the first to extend the TAM framework to a machine translation tool like DeepL, addressing a gap in MT research by examining the role of usage frequency. The findings offer both theoretical and practical significance, providing insights into how MT tools can be more effectively integrated into EFL learning.