Network reconfiguration (NR) is one of the most effective methods to reduce line power loss in the distribution system, which causes higher losses than the other parts of the power system. This paper proposes a modified symbiotic organisms search (MSOS) algorithm to solve the NR problem. For the purpose of enhancing the effectiveness of MSOS, the mutualism and parasitism strategies of the original symbiotic organisms search (SOS) have been modified to create better new solutions. In the mutualism strategy, the so-far best solution is updated immediately as soon as new solutions are created. In the parasitism strategy, the update is only implemented for the first half of control variables, whereas another half still remains unchanged. The comparison results between MSOS and SOS on twenty-five benchmark functions and different scales of test distribution systems with 14, 33, 69, and 119 nodes show that the improvement level of MSOS over SOS is significant with higher success rates and better quality of gained solutions. Similarly, MSOS also reaches better results than other methods in the literature. Consequently, MSOS can be a favorable method for determining the most appropriate configurations for the distribution systems.