Self-assembly of nanoclusters (NCs) is an effective synthetic method for preparing functionalized nanomaterials. However, the assembly process and mechanisms in solutions still remain ambiguous owing to the limited strategies to monitor intermediate assembled states. Herein, the self-assembly process of amphiphilic molecule 4POSS-DL-POM (consisting of four polyhedral oligomeric silsesquioxanes, a dendritic linker, and one polyoxometalate) by evaporation of acetone in a mixed acetone/n-decane solution is monitored by time-resolved synchrotron small-angle X-ray scattering (SAXS). Scattering data assessments, including Kratky analysis, pair distance distribution function, and model fitting, track the self-assembly process of 4POSS-DL-POM from a fractal network to compact NCs, then to core-shell NCs, and finally to superlattice structure. The calculated average aggregation number of a core-shell NC is 11 according to the parameters obtained from core-shell model fitting, in agreement with electron microscopy. The fundamental understanding of the self-assembly dynamics from heterocluster into NCs provides principles to control building block shape and guide target aggregation, which can further promote the design and construction of highly ordered cluster-assembled functional nanomaterials.