A novel approach for crack identification based on jointly time-frequency analysis is presented in the paper. A bilinear stiffness model for the breathing crack was introduced to represent the nonlinear dynamics of a cracked beam. The nonlinearity of the dynamic responses due to the crack opening-closing is used to identify the occurrence of the crack. The Wigner-Wille distribution technique is applied to analyze the response signals and the instantaneous frequency is extracted as damage-sensitive feature. The numerical simulations of a breathing crack model were carried out to validate the possibility and effectiveness of the proposed approach. The effects of crack severity and sampling frequency on crack identification were also studied in the simulations respectively. The results show that the proposed method can effectively identify the crack with slight severity without any baseline model or data, and the better the identification obtains as the larger the sampling frequency. The study demonstrates that the proposed approach by using of jointly time-frequency analysis is a promising technique for crack identification.