A method for partitioning airspace into smaller regions based on a peak traffic-counts metric is described. The three setup steps consist of 1) creating a network flow graph, 2) creating an occupancy grid composed of grid cells of specified size for discretizing the airspace and 3) assigning the grid cells to the nodes of the network flow graph. Both the occupancy grid and the grid cell assignment to nodes are computationally realized using matrices. During the run phase of the method, the network flow graph is partitioned into its two sub-graphs and these two sub-graphs and then partitioned into their two sub-graphs, and so on till a termination criterion is met. Weights of the sub-graphs are computed by summing the number of aircraft in each grid cell associated with the nodes of the sub-graphs at each time instant. This process is accomplished by using the occupancy and assignment matrices created during the setup step. The final weight is obtained as the maximum count over a time period. Spectral bisection is then used to split the sub-graph with the maximum weight into its two sub-graphs. Recursive application of the spectral bisection method and weight computation results in the final set of sub-graphs. The grid cells associated with each sub-graph then represent the geometry of the associated sector. Results of sectorization of the airspace over the continental United States are provided to demonstrate the merits and the limitations of the method. The weighted-graph technique created larger sectors in regions of light-traffic and smaller sectors in regions of heavy-traffic. Peak traffic-counts in the sectors were found to be within the range of the Monitor Alert Parameters specified in the Enhanced Traffic Management System.