The paper addresses the challenge of maintaining uninterrupted metaverse experiences for Vehicle Twins (VTs) in vehicular edge metaverses by proposing a cross-metaverse empowered dual pseudonym management framework. The framework leverages blockchain technology to enhance management efficiency and data security for pseudonyms, and introduces a Degree of Privacy Entropy (DoPE) metric to assess privacy levels. A Multi-Agent Deep Reinforcement Learning (MADRL) approach is employed to optimize pseudonym generation strategies. The proposed scheme is evaluated through numerical results, demonstrating its high efficiency and cost-effectiveness in vehicular edge metaverses.The paper addresses the challenge of maintaining uninterrupted metaverse experiences for Vehicle Twins (VTs) in vehicular edge metaverses by proposing a cross-metaverse empowered dual pseudonym management framework. The framework leverages blockchain technology to enhance management efficiency and data security for pseudonyms, and introduces a Degree of Privacy Entropy (DoPE) metric to assess privacy levels. A Multi-Agent Deep Reinforcement Learning (MADRL) approach is employed to optimize pseudonym generation strategies. The proposed scheme is evaluated through numerical results, demonstrating its high efficiency and cost-effectiveness in vehicular edge metaverses.