Aimed at the problem that computation measure is comparatively heavy by the joint estimation algorithm of information source azimuth and polarization state of conformal array antenna, a forward recursion method of multistage wiener filter for cylindrical conformal array antenna is derived, and an efficient joint estimation algorithm of information source azimuth and polarization parameters based on fast subspace estimation is presented. Taking the signal of a desired signal as known conditions, this algorithm realizes the signal subspace and noise subspace quick estimate through multistage wiener filter forward recursive, avoiding the covariance matrix estimation and eigenvalue decomposition, greatly reducing the computational cost for joint estimation algorithm, and reducing amount of calculation from the original O(N3+N2L) to O(N2L). The computer Monte Carlo simulation experiment shows that the algorithm can guarantee the estimation accuracy and simultaneously reduce the complexity of the algorithm. When the SNR is greater than 10 dB, the algorithm has approximately the same estimation accuracy compared with the existing algorithm. And the algorithm is valid.