The expansion of digital technology and algorithm-driven platforms has transformed the ways Generation Z encounters and interprets religious knowledge, creating new challenges for Islamic education. Although digital media broaden access to religious information, they also encourage the spread of fragmented narratives, misinformation, and uncritical patterns of religious learning. This study aims to examine the crisis of religious literacy experienced by Generation Z in the era of algorithmic Islam and to develop a deep learning-based Islamic education model capable of strengthening critical religious literacy. Using an Educational Design Research (EDR) approach, the study was conducted from March to May 2026 in four Islamic educational institutions in Banten Province, Indonesia. Twenty-five participants, including school principals, Islamic Religious Education teachers, and parents, were selected through purposive sampling. Data were collected through semi-structured interviews, classroom observations, focus group discussions, document analysis, expert validation, and limited model implementation. The findings indicate that existing Islamic Religious Education remains predominantly transmission-oriented and has not adequately addressed the epistemological complexities of digitally mediated religious environments. The study subsequently developed the Deep Learning-Based Critical Religious Literacy (DL-CRL) Model, which comprises five interconnected dimensions: critical inquiry, digital religious verification, reflective dialogue, collaborative knowledge construction, and transformative religious engagement. Limited implementation demonstrated improvements in students’ ability to evaluate religious information, verify digital sources, and construct evidence-based religious arguments. The proposed model offers an integrative pedagogical framework that connects Islamic educational values, deep learning, and digital religion studies to support religious literacy among Generation Z.