Spin distribution of dark matter halos in a cosmological N-body simulation is well fitted by a log-normal distribution, but the origin of the log-normal like shape is still unknown. To understand the evolution of spin and the origin of spin distribution, we have studied the change of the angular momentum of simulated halos through their merging histories. First, we traced merging histories of the dark matter halos and measured the probability distribution of the angular momentum changes from a series of simulations. We were able to fit the angular momentum changes with the Gaussian distribution in spaces of M, spin, ${Delta}M$. Using the simulated merging trees and the distribution of angular momentum changes during the merging events, we can recover the spin distribution of halos over the various mass scales.