The rational planning of school location is an important way to optimize the allocation of educational resources, improve the efficiency of school running and realize the balanced development of education. The accessibility must be considered in optimizing the location of rural schools, which is an important difference from the location of urban schools. Many scholars have studied the location problem of rural schools, but most of them have neglected the impact of transportation network conditions on the location of schools. On the basis of previous studies, this study will consider the impact of traffic network on the optimization location of primary schools in mountainous environment. The object of this study is to minimize the total transportation costs for students, construction costs for new schools, construction and upgrade costs for roads on a traffic network. We regard a set of villages as demand nodes exists in a geographical region, a set of roads as transfer links. The road links in the network contain existing and new potential road links. A set of schools exists in the region and it is clearly desired to locate a set of new schools, to construct new road links, to improve the existing road links such that the total investment costs (including the travel costs for students, construction costs for school facilities, construction and upgrading costs for roads) are minimized. Thus, a multi-objective optimization model for this problem is proposed. The multi-objective facility location-network design model constructed in this paper is an extension of FLNDP problem, and also belongs to NP-hard problem. In order to facilitate the model application, this study use geographic information science method, take Visual Studio. Net 2010 as the development platform, use C# language, MATLAB language and ArcGIS Engine 10 component library to develop a location optimization system of rural primary and secondary school based on transportation network. Finally, the system is applied to the optimization of the primary school location of a town in Guizhou province. Aiming at the location optimization problem of rural primary, this paper considers three scenarios (only consider the existing road, assume that the road can be upgraded and the road can be built or upgraded) which are different from the problem of urban primary school location. In geographic calculation, this paper explores an improved multi-objective simulated annealing algorithm to determine the best location of new schools, as well as the construction of new roads and the upgrading of existing roads.