We perform statistical wavelet analysis of the Main Belt asteroids, seeking statistically significant asteroid families. The goal is to test the new wavelet analysis algorithm and to compare its results with more traditional methods like the hierarchic clustering. We first consider several 1D distributions for various physical and orbital parameters of asteroids. Then, we consider three bivariate distributions for the three orbital parameters (a, e, i) taken pairwisely. The full 3D analysis of this space is not available here, but based on the 2D results we perform a disentangling of overlapped 2D families and derive a total of 44 3D families with confirmed statistical significance.