Abstract: Psychological crisis intervention faces significant methodological challenges in dynamic monitoring, precise prediction, and strategy optimization. Traditional methods often struggle with the inherent complexity and personalized needs of psychological states. This paper explores the introduction of Parallel Intelligence theory into psychological crisis intervention and proposes a conceptual framework for Parallel Psychological Crisis Intervention. Based on the ACP approach (Artificial Systems, Computational Experiments, and Parallel Execution), this framework aims to achieve state modeling by constructing artificial psychological systems, derive psychological states and evaluate intervention strategies through computational experiments, and optimize strategies via virtual-real interaction through parallel execution. It is designed to provide a preliminary framework for computable, experimental, and iterative psychological crisis intervention. As a proof-of-concept prototype, this paper develops PsyRescueGPT, a Large Language Model-driven multi-agent system. This system covers the entire process from monitoring, analysis, and prediction to strategy generation, evaluation, and parallel execution/optimization, providing a potential technical path for transitioning psychological crisis intervention from experience-driven to computation-driven paradigms.

Key words: psychological crisis intervention, parallel intelligence, ACP, Large Language Models, multi-agent systems, personalized intervention