As a developing emerging market, my country's financial market is not only credit risk, but other risks such as market risk will also gradually increase with the development of the financial market. Therefore, the research on financial risk management methods is of great significance to the current and future my country's financial innovation and investment decision-making by investment institutions. As an indispensable part of the securities investment process, securities investment analysis is the basis for making investment decisions, and it occupies a very important position in the investment process. The biggest feature of support vector machine is that it has changed the traditional empirical risk minimization principle, and proposed for the structural risk minimization principle, so it has a good generalization ability. This paper studies and analyzes the optimization of securities investment portfolio based on support vector machines.