A method to achieve the sparsification of the impedance matrix is proposed when the higher order hierarchical vector basis functions and the maximally orthogonalized higher order vector basis functions are applied to electromagnetic integral equations.Both the theory analysis and the numerical experiment demonstrate that this method will decrease the memory requirement for storing the impedance matrix and the CPU time consumed in solving the matrix equation when using an iterative method.Moreover

some numerical results have been given to study the different effects on the requirements for the memory and CPU time caused by different thresholds.