For improving target's location accuracy and fully utilizing redundant information in complex bistatic radar systems, a new target location algorithm based on Gauss-Newton iterative is proposed. The key features of the algorithm are as follows: All the observation data from transmitting station and receiving station are adopted to compose a nonlinear least square equation, the location estimation resolved by a measurement subset which has higher accuracy among all subsets is utilized as an initialized value, which is close to the truth value. The idiographic flow about iterative algorithm is constituted and the strategy of changing the iterative step size is introduced so that the objective function is dynamically declined. Moreover, expressions of location error covariance matrix are deduced and the target's location accuracy is analyzed. Simulation shows that by using the algorithm, the iterative convergence and veracity of target location are improved. This algorithm, as compared with the SWLS algorithm, has more accurate target location solution. In this way, the location accuracy in the whole detected area is greatly enhanced and the capability of target location is optimized.