A new constrained algorithm based on the non-holonomic natural gradient with variable step-size is proposed, which can deal with the instantaneous blind source separation problem in the non-stationary environments. In order to improve the convergence speed, an adaptive means is adopted to the constrain factor, which can do the adjustment adaptively according to the separation situation. At the same time, the variable step-size based on the gradient of cost function is also applied, so that the proposed algorithm has better tracking performance. The computer simulation results show that the proposed algorithm improves the convergence speed as well as separates source signals effectively without producing serious steady state error in the non-stationary environment.