Chaos theory has provided a new pattern for understanding the interacting process of risk factors in system analysis. Software project risk system is a nonlinear system and so it also can be seen as a chaos system. This paper introduces chaos theory into the analysis on software project risk systems to find the evolution law of risk system From the point of chaotic dynamics a dynamic model based on three main parameters for software project risk system on the basis of rational assumption is established. According to judging principles of Lyapunov exponent on chaotic characteristics and the calculation of Lyapunov exponent of software project risk system model, this paper proves that the software project risk system is a chaotic one, which lays the foundations for the related risk research based on the chaos theory. Then it discusses chaotic characteristics such as initial sensibility, inner randomness and strange attractor in software project risk system, which gives guidance for risk control. It also brings a new thought for the study of software project risk management from the point of system.