As we all know, most of the practical problems when reduced to linear programming problem have the constraint set which is the system of linear inequalities. We introduce a new algorithm is an algorithm that directly solves the primal standard linear programming problems. This algorithm has other advantages over the Simplex Method and Interior point Methods: No need to know an initial feasible point of the original problem, no need to add slack variables are required to return the original problem to a canonical form. We have built a computer sample program for this algorithm to solve primal standard linear programming problems with size any on the Matlab environment. The experimental calculation results with random data show that the number of iterations and calculation time according to this new algorithm compared to the Simplex algorithm is much less and noticeably faster. We hope that it will be one of the effective algorithms for solving medium and large size linear programming problems on any common computer today.