This paper reports the development of a two-level optimization methodology to help design a tri-generation system for a given district which satisfies the heating, cooling, and hot water demands and at the same time, minimize the annual total costs and CO2 emissions. An optimization methodology is proposed and tested on a virtual district with eight buildings where three of them can host the district technologies including heat pump, gas engine, and lake cooling. Within the building, some backup technologies may be implemented including an air/water heat pump, a water/water heat pump, a boiler, and electric chillers. Analysis of the Pareto optimal frontier results in several distinct groups of configuration based on the selected district conversion technologies and their capacities. Solution to the sub-problems including design and operation of the district energy system is carried out by applying a Mixed Integer Programming (MIP) technique. Several different clusters are defined and studied regarding the cost and CO2 emission. A reference configuration is defined for the purpose of comparison in which electricity is supplied by the grid, heating and hot water by a boiler, and cooling by an electric chiller. Compared to this configuration, the best solution with respect to CO2 emissions causes 59% emission and 75% cost of the reference configuration. In this case, 53% of the total cost is associated with the initial investment cost while the rest 47% is associated with the operational cost. The optimal configuration with respect to the annual costs causes 86% more emission than the reference configuration and 38% less annual costs. In this case, 22% of the total cost is associated with initial investment cost while 78% of the total cost is associated with the operational cost. Implementation of a two-pipe system instead of a four-pipe system results in nearly 5% reduction in total annual cost.