AbstractThe coupled dipole method, as originally formulated by Purcell and Pennypacker [3], is a very powerful method to simulate the elastic light scattering from arbitrary particles. This method, however, has one major drawback: if the size of the particles grows, or if scattering from an ensemble of randomly oriented particles has to be simulated, the computational demands of the coupled dipole method soon become too high. This paper presents two new computational techniques to resolve this problem. First the coupled dipole method was implemented on a massively parallel computer. The parallel efficiency can be very close to 1, implying that the attained computational speed scales perfectly with the number of processors. Second, it is proposed to reduce the computational complexity of the coupled dipole method by including ideas from the so‐called fast multipole methods (hierarchical algorithms) into the coupled dipole method. In this way calculation time can be decreased by orders of magnitude.