| Since the21st century, China’s IT industry has achieved sustained and rapid development, and gradually become the strategic, fundamental and guiding industry of national economy. IT industry has those features like with high-yield, high-risk, high-tech, high investment and so on, which determines its capital structure is different from other companies. Compared to western countries, our country’s information technology industry is not mature enough. There are many problems in capital structure. Domestic and foreign research on the capital structure of listed companies in information technology industry is rare, but also lacks the application of nonlinear model research in this field. Those problems restrict the long and stable development of information technology. Therefore, considering the situation of our country, it is a very important and realistic issue to analyze the capital structure of IT enterprises. Besides, we should find a way to optimize the capital structure of those enterprises.On the basis of capital structure theory, and according to other scholars’research and combined with the reality of our country, this paper analyze the current situation and capital structure of IT industry, to confirm the determines of capital structure. On this basis, by using Matlab statistical software,136information technology listed companies that published by CSRC in2014are chosen as samples to set up BP model. Besides, this paper set up multivariate linear regression model as robust analysis to verify the nonlinear fitting effect of the model, so as to explore a more appropriate model.Through empirical analysis, we found that DEBT, profitability are negatively correlated with capital structure; while growth, scale, manage and tang are positively related with capital structure; but the impact of growth is not significant. Through robust analysis, we found that BP neural network model has good fitting capability. At last, on the basis of empirical analysis, this paper presents some advices on optimizing capital structure of IT industry. Hoping can provide some theoretical basis to help managers adjust capital structure flexibly. |