The selection of 10 kV mid-voltage connection modes, which is the key to planning and transformation of distribution network, is affected by load density, investment, stability, etc. Under the condition of different load density, the connection unit capacity determines the required quantity. Considering the impact of many factors on investment cost and connection mode reliability, and combining the comprehensive function of user’s power outage loss, the economy and reliability can be transferred between each other. The line parameters which impact the economy and reliability of connection mode are adopted as the initial group, the genetic algorithm is used, and the power supply’s lowest total cost is taken as objective, to obtain the Pareto optimal solution set of any load density. Thus the GA-Pareto optimization model is established for mid-voltage connection mode. The model has been applied in a city in the central China, which supplies theoretical basis for the construction of distribution network of this city. This work is supported by National Key Research and Development Program of China (No. 2017YFB0902800), Science and Technology Key Project of Henan Province (No. 182102210112) and Basic and Frontier Technology Research Project of Henan Province (No. 132300410185).