This research project is expected to develop and test an AI-based Learning Management System (LMS) contents which can promote Self-regulated learning (SRL) with secondary school mathematics students. Considering that conventional LMS have shortcomings in relation to adaptive learning, and given the increasing significance of SRL, this study bridges that gap by proposing new AI-based LMS for mathematics at high school level. Utilizing a Research and Development model combined with an embedding quasi-experiment design of 94 students, the project developed AI-enhanced LMS-based content that included; personalization, adaptive difficulty, AI tutor intervention, and asynchronous administration. The findings reveal that the AI-based LMS content was effective on math students N-Gain scores compared to traditional instruction learning but with small effect sizes. In addition, students employed the AI LMS at moderate-high SRL levels in most of its components (i.e., goal setting, self-evaluation and help seeking). What is new in this work is the development per se (using AI) of LMS content for mathematics to help support SRL, while also taking into account multi-device and user-friendly considerations. Operationally, this AI-LMS system provides a student-centric and flexible learning environment that can help students develop self-regulation of learning to promote engagement and achievement in mathematics. The unique contribution of this study is a validated AI-oriented LMS model that closes the distance between AI promise and SRL requirements in secondary mathematics education.