One of the most important standards to measure a cable-stayed bridge safety and quality is whether the permanent load inner force is reasonable. A reasonable completed status of cable-stayed bridge should be that the bending stress is small and even. To an existing cable-stayed bridge, there must be a group of cable force which could optimize the bridge to a reasonable state. The optimization of cable force is the permanent theme to cable-stayed bridge.Several methods of determining reasonable cable force of cable-stayed bridges have been primarily analyzed and studied in this paper, and bending strain Energy based genetic algorithm was adopted as aim to define the initial cable force of completed status of PC cable-stayed bridge .PC cable-stayed bridge is highly redundant complex structure; the completed status of cable-stayed bridge is closed relative to the construction process. How to determine the cable force of construction process is the key during the construction. In the paper, several simulation methods like forward analysis, backward analysis, approach iteratively etc are detail discussed. Then a cable force calculation model is established in which the sum of squares of main beam datum mark's relative error is defined as target function, with main beam moments as constraint condition and cable force as optimization variable.The reasonable completed status of cable-stayed bridge and construction state need construction to approach. So, the content and principle of construction control of cable-stayed bridge were discussed and analyzed, and from the point of cybernetics the opening loop control, closed loop control, and adaptive control were summarized. Then the adaptive control construction model of PC cable-stayed bridge based data infusion is established, in which the three space model and the plain model, nonlinear and linear are compared to analyze the construction state. Meanwhile, a method of parameter identification based on combining Artificial Neural Networks (ANN) and Genetic Algorithm (GA) is applied successfully. And at last, the gray theory and BP Neural Networks are applied for construction error control. The result shows the