General form-finding problems of cable-supported bridges are established based on a design scenario in which rigidly fixed starting control points must be given as necessary design constraints prior to independent analysis of any of its cable subsystems. This article presents a form-finding method to address a new case, in which the starting control point serves as an intermediate, flexibly variable connection, to couple two related cable subsystems in a multi-nonlinear environment for the target configuration under dead load (TCUD) of a novel type of spatial self-anchored hybrid cable-stayed suspension (HCSS) bridge. A two-layer framework is proposed by integrating finite element analysis (FEA) and analytical formulas with optimization algorithms to form a self-regulated interactive analysis among subsystems in an iterative manner. The outer layer seeks to achieve self-equilibrium of the global system under the control information of the TCUD, while the inner layer optimizes the subsystems in terms of the initial tensions in the main cables, stay-cables, branches and hangers to obtain a rational mechanical state of the bridge. Then, the TCUD and the intermediate starting control points are determined. To achieve computational stability, a high-performance accelerated Steffens-Newton (ASN) differential algorithm is developed for the shape finding of the cable-hanger subsystem, whereas the non-dominated sorting genetic algorithm (NSGA-II) is adopted as a multiobjective optimizer for TCUD optimization of the other subsystems. The proposed framework is applied to a real-scale self-anchored HCSS bridge, and its validity and performance are demonstrated by comparison studies with a non-optimal scheme and in-field test data.