This paper discusses the design of complex coefficients FIR filters in the complex Chebyshev sense. One way to solve this problem is to first formulate the complex Chebyshev approximation problem into finite linear programming through discretization, and then simplex method follows. But it has been thought that this method involves too much computation cost, converges slowly and it is applicable only for designing real coefficients FIR filters. Here a new algorithm is formed, which is a variant of the simplex method. It exhibits good convergence, low computation effort and can be used to design complex coefficients FIR filters. It proves equal to the algorithm formed from the dual of semi-infinite linear programming. An example is presented in the end.