Abstract: Bird targets exhibit characteristics of low observability and high maneuverability, and the micro-motion parameter estimation of the bird targets is crucial for their identification. Based on the radar echo model of bird target, the flapping frequency is first estimated using the autocorrelation method and a method based on matching pursuit for estimating the wing length of bird targets is proposed. The radar echoes are directly processed and a dictionary based on the bird target signal model can be constructed. Considering that the most matching atom in the dictionary corresponds to a different wing length, the weighted sum of the wing length values corresponding to the optimal matching atoms from each iteration is proposed as the estimated wing length of the bird target. The weighting coefficients are determined by the degree of matching between the matching atoms and the received signal. This method offers high parameter estimation accuracy and good noise resistance, but it has low computational efficiency. To address this issue, the time-frequency concentration index is also applied to bird wing length estimation. A rotation operator is derived and constructed based on the micro-motion characteristics of bird targets, and wing length estimation is achieved based on the concentration of the time-frequency spectrum. This method improves computational efficiency at the cost of a slight loss in the parameter estimation accuracy. In engineering applications, the proposed methods can be selected according to different application backgrounds to effectively estimate the upper and forearm lengths of bird targets.