Understanding the mechanical properties of soft materials such as stress-strain behavior over a large deformation domain is essential for both mechanical and biological applications. Conventional measurement methods have limited access to these properties because of the difficulties in accurately measuring large deformations of soft materials. In this study, we optimized digital image correlation (DIC) method to measure the large-strain deformations by considering referencing scheme and frame rate. The optimized DIC was utilized to estimate strain in characterizing the stress-strain behavior of a polydimethylsiloxane (PDMS) elastomer as a model soft material. A series of comparative experimental studies and finite element analysis were performed; they indicated the advantages of optimized DIC over conventional methods such as robustness to slip, insensitivity to boundary conditions, and the ability to yield consistent and reliable results. These advantages enabled the optimized DIC to perform an in-depth analysis of the behavior of soft materials at large strain domain. An empirical constitutive equation to describe the large stress-strain behavior of PDMS was proposed and verified by finite element simulations that show excellent agreements with experimental results. POLYM. ENG. SCI., 52:826-834, 2011. (C) 2011 Society of Plastics Engineers