The rise of social media has reshaped communication, generating new linguistic phenomena. Malaysian Chinese youth, immersed in Chinese family traditions and Malay and English social environments, leverage platforms like TikTok to construct their self-identity. Despite this complexity, the multilingual and multimodal dimensions of such practices remain understudied. This study, employing multimodal discourse analysis supported by systemic functional linguistics (SF-MDA), addresses this gap by investigating: What systemic functional linguistic and multimodal features do Malaysian Chinese youth employ on TikTok to construct their multicultural identity? The findings reveal firstly that two distinct patterns emerge: English-dominated multilingual practices, and code-mixing phenomenon (English-Chinese, or English-Malay mixing). Secondly, Malaysian Chinese youth online communication transcends Malay, English, and Chinese, integrating visual components such as signs, images, body language, and facial expressions. Non-linguistic visual aspects are utilised more often than non-linguistic auditory elements. And thirdly, five multimodal strategies are identified: visual-autonomous, visual-dominant, auditory-autonomous, auditory-dominant, and co-dominant, with co-dominant being the most frequently employed.