This talk is concerned with the simultaneous reconstruction of obstacle and spatially varying impedance with near field measurements by a regularized Gauss-Newton iterative inversion algorithm. The initial of the boundary is given by a sampling method with a modified indicator function. We extend a sampling method with respect to Dirichlet and Neumann boundary conditions to impedance boundary case. Theoretical analysis based on the factorization method and numerical experiments show that our method has excellent result on the reconstruction of the boundary. By extracting the points with large indicator values and clustering them to obtain the points near the exact boundary, the initial curve can be obtained by fitting. Then a sufficiently accurate iterative result is given starting from the fitted initial curve and a constant initial impedance.