Functional nanoparticles (NPs) have gained significant attention as promising applications in various fields, including sensor, smart coating, drug delivery, and more. Here, a novel mechanism assisted by machine‐learning workflow is proposed to accurately predict phase diagram of NPs, which elegantly achieves tunability of shapes and internal structures of NPs using self‐assembly of block‐copolymers (BCP). Unlike most of previous studies, onion‐like and mesoporous NPs in neutral environment and hamburger‐like NPs in selective environment are obtained. Such novel phenomena are obtained only by tailoring the topology of a miktoarm star BCP chain architecture without the need for any further treatment. Moreover, it is demonstrated that the BCP chain architecture can be used as a new strategy for tuning the lamellar asymmetry of NPs. It is shown that the asymmetry between A and B lamellae in striped ellipsoidal and onion‐like particles increases as the volume fraction of the A‐block increases, beyond the level reached by linear BCPs. In addition, an extended region of onion‐like structure is found in the phase diagram of A‐selective environment, as well as the emergence of an inverse onion‐like structure in the B‐selective one. The findings provide a valuable insight into the design and fabrication of nanoscale materials with customized properties, opening up new possibilities for advanced applications in sensing, materials science, and beyond.