The present invention provides a super-large scale uncontrolled regional network robust adjustment method and system. Because super-large scale optical satellite image regional network adjustment relates to combined processing of lots of images, a large number of parameters are required to solve, and the parameters to be solved may be shocked to be difficult for convergence in the solution process due to too few constrained conditions in the condition without control or with little control. A group of virtual control points is generated aiming at each image to be adjusted, and multi-picture images with overlapping are subjected to automatic matching of images to generate connection points; and added model parameters are solved by combing adjustment according to the virtual control points and the connection point data, the image RPC model of added image space transformation parameters is obtained, and unknown parameters are subjected to adjustment solution. Virtual control points are taken as weighted observed values and add adjustment to perform combined solution, so that the effects of the connection points in the uncontrolled free net adjustment mode may be effectively restrained, the problem may be solved that the uncontrolled free net adjustment solution is not convergent, and therefore the reliability of the adjustment result is greatly improved.