Abstract: The rapid development of blockchain technology has propelled the widespread application of smart contracts, but their security issues have also become increasingly prominent, emerging as a key factor that constrains the development of blockchain technology. For example, the well-known The DAO incident resulted in approximately 60 million in economic losses. This paper conducts a detailed analysis of the security risks at various stages of smart contracts, including design and development, compilation and deployment, execution, and maintenance management, revealing vulnerabilities and threats present in each stage. A series of secure programming standards and best practices are proposed, including design patterns, permission management, and input validation, to mitigate the security risks of smart contracts. A comprehensive smart contract security evaluation framework is constructed, integrating static and dynamic analysis techniques and developing automated detection tools to provide a standardized assessment process for smart contract security. Through empirical studies on Ethereum, Polkadot, and Binance Smart Chain, the effectiveness of the evaluation framework is verified, and targeted improvement suggestions are proposed. By looking ahead to the future development directions of smart contract security evaluation, including intelligent and automated evaluation technologies, cross-chain evaluation capabilities, real-time monitoring and early warning systems, standardization and international certification, as well as community-participatory crowdsourced evaluation models, this paper aims to drive the continuous improvement and optimization of smart contract security.