This research explores Human–GenAI collaboration in complex problem-solving tasks in online learning environments. Grounded in Self-Regulated Learning (SRL) and Complex Problem-Solving (CPS), the study examines interaction patterns, problem-solving strategies, and collaboration approaches while solving complex problems using GenAI tools. The stage-wise methodology includes conducting interviews and surveys, complex problem-solving tasks, and an intervention to assess learning outcomes. Data collection involves ChatGPT interaction logs, screen recordings, and post-task interviews, which are analyzed using Ordered Network Analysis (ONA) and Epistemic Network Analysis (ENA). The findings aim to inform the development of a Collaborative Complex Problem-Solving (CCPS) framework and provide pedagogical guidelines for effective Human–GenAI collaboration.