To deploy a search robot rescue system in a disaster struck area, we have been researching and developing an integrated wired and wireless hybrid network for stably transmitting real-time control and measurement data including video signals from a group of high-speed search robots in a closed space. In this paper, we propose a hybrid network model for search robot rescue system operating in a basic closed space. We create two scenarios based on the search robots' movement. One is the "simultaneous" scenario, in which all search robots depart all at once and the other is the "sequential" scenario, in which search robot departs in order. We analyzed the performance of the hybrid network model in these two scenarios by using a network simulator. We then apply these network models to a typical closed space with turns. The simulation results show that the network performances are strongly related to the position of search robots and the sequential scenario is superior to the simultaneous scenario. The shape of the closed space affects very little when the same random seed was used and it is also shown that the network performances are related to the random seeds that create simulated movement of search robots. The simulation results will be useful as a guideline to the sequence of deploying robots in an actual rescue system.