A prediction method of the no-firing current of electric explosive device was studied.The electric-thermal parameters of the electric explosive device such as resistance and heat loss coefficient were measured with the non-destructive transient pulse test system, then the heat loss coefficient was redeemed using the no-firing current measured by the Bruceton method.The no-firing current of the electric explosive device was predicated using the radial basis function(RBF) neural network.The results show that the predicted result is consistent with that measured with the firing validated test system, and the electric explosive device with the larger predicted value owns the larger firing probability.The mean predicted current equals to the firing current measured by the Bruceton method.