Process security is a newly pronounced issue facing the chemical process industry in the post 11 September era. Traditional safety is no longer sufficient for a chemical plant; it must also be secure. However, systematic and effective quantitative methodologies for process security analysis are necessary. To address this issue, the gamma-analysis method was introduced very recently by Uygun et al. (2003) as a process security analysis framework. By that method, a process security problem need be formulated as a minimum-time (to reach disaster) control problem. The method combines Pontryagin's minimum principle (with some modifications) with a discretization scheme to transform the security problem from a single dynamic-optimization problem to multiple static optimization problems, hence solving the process security problem without extensive system simulations. In this work, an improved gamma-analysis method is introduced, which is featured by its first-order approximation to the time derivative function of a system model, as compared to the zero-order approximation during model discretization in the original implementation. This improvement can significantly reduce the number of optimization problems needed, thereby reducing computational requirements in on-line application. A reaction runaway example is studied to demonstrate the efficacy of the improved method, demonstrating that a 10-fold reduction in computational load can be achieved.