Abstract: Bell-LaPadula model(BLP) model is the most classical multilevel computer security model. Based on the research of BLP model, this paper analyzes the main defects in BLP model and different improving methods, puts forward a principle or “read the nearest and write the nearest”, improves the BLP model more effective security and integrity.

Key words: Bell-LaPadula model(BLP), Model, Security

摘要： BLP模型是最为经典的计算机多级安全模型。在进行BLP模型研究的基础上，对其存在的主要安全缺陷及不同的改进方案进行分析，提出了“最近读、最近写”的原则，较为有效地改进了BLP模型的安全性和完整性。

关键词: BLP模型, 模型, 安全