Regarding the fact that space target recognition is a challenging research problem, this paper proposes a space target recognition algorithm based on the two-dimensional wavelet transform. Firstly, the two-dimension wavelet transform of ISAR image of space target was proceeded. Then singular value feature was extracted from approximate and detail parts. Finally, the radial base function(RBF) neural network was used to classify space targets. Computer simulation shows that such an algorithm achieves good recognition effect.