Recently, multi-hop fact verification has become a crucial research direction, aiming to verify claims by reasoning over multiple pieces of evidence. Existing methods primarily rely on multi-agent collaboration to decompose fact verification into specialized subtasks. However, these methods face two critical limitations: (1) agents may perform individual subtasks without sufficient awareness of the global verification objective, causing their reasoning to deviate from the intended direction; and (2) conflicts between parametric knowledge and the provided evidence may undermine evidence-grounded reasoning and lead to incorrect verdicts. To address these challenges, we propose ReflectFact, a novel self-reflective agent framework for multi-hop fact verification. ReflectFact introduces three key tasks: Explicit Reasoning Path Planning, Evidence-Drift Verification, and Reasoning Reflection Verification. Through these tasks, ReflectFact can effectively remedy the comprehension and reasoning defects of existing methods, achieving state-of-the-art performance. Extensive experiments on HOVER and EX-FEVER demonstrate that ReflectFact outperforms the strongest baseline by 3.32% and 2.78% on the two datasets, respectively. Blogger's Review: ReflectFact provides a new perspective for multi-hop fact verification by utilizing self-reflective agents to improve comprehension and reasoning abilities, which is an important direction for future research.