With the development of international maritime trade, the actual number of surveillance cameras in smart ports is increasing. In response to the widely valued early warning detection and prevention system, the security monitoring optimization system based on image recognition technology can effectively cope with the increasingly complex real-time monitoring system. This article studies the design of an improved security monitoring system using image recognition technology. After preprocessing and feature extraction of image data, combined with the evaluation of security supervision engineers and testing samples of important port areas, the following conclusions are drawn through questionnaire survey and simulation experiment analysis: Within the framework of intelligent security systems, security monitoring systems that utilize image recognition technology surpass traditional monitoring methods in accuracy across all important area samples, with an average improvement of 6.4%. They also yield favorable practical results in system stability, with an average satisfaction rating of approximately 91.2 points. This research offers a more efficient and precise monitoring method for port security management, substantially enhancing the safety of goods, equipment, and personnel, while also reducing the costs and error rates associated with manual monitoring.