The need to secure software application in today’s hostile computer environment cannot be overlooked. The increase in attacks aimed at software directly in the last decade and the demand for more secure software applications has drawn the attention of the software industry into looking for better ways in which software can be developed more securely. To achieve this, it has been suggested that security needs to be integrated into every phase of software development lifecycle (SDLC). In line with this view, security tools are now used during SDLC to integrate security into software applications. Here, we propose a neural network based security tool for analyzing software design for security flaws. Our findings show that the trained neural network was able to match possible attack patterns to design scenarios presented to it. With the information on the attack pattern identified, developers can make informed decision in mitigating risks in their designs.