Proposing a comprehensive model to predict the fatigue life of composite structures under dynamic loading has always been a challenging problem. In this study, an attempt has been made to develop a polynomial model by utilization of experimental data together with singular value decomposition technique to estimate the fatigue life of composite materials. The model is based on a nonlinear mathematical function with effective parameters related to fatigue phenomenon in a nondimensional form. For this, four effective parameters such as cyclic stress amplitude, static strength, stress ratio, and fiber angle have been considered in modeling process. To evaluate the capability of this model a comparison has been made with some available experimental results where a good agreement is achieved. So, the obtained results demonstrate that despite system complexity and inability to extract an analytical model for estimating the fatigue life of composite materials, singular value decomposition method could be known as a useful and powerful tool for modeling.