In applications like communication with a wide-ranging coverage, navigation, and Earth observation, large-scale satellite constellations mark the next frontier in the space community. This research presents an optimal intelligent satellite constellation initialization framework using minimum-energy motion planning. Combining traditional orbital mechanics with supervised learning, it addresses the placement of satellites in designated orbits to achieve efficiency in deployment by analyzing orbital maneuvers. The Hohmann and non-Hohmann transfers are analyzed for their applicability in moving satellites between orbital configurations. The study uses MATLAB simulations to validate the framework, employing neural networks to predict optimal target acquisition while accounting for the initial orbit and transfer trajectories, treated as input-output pairs of initial and target orbital parameters. Results demonstrate the efficacy of the proposed framework in determining minimum-energy paths while accounting for autonomy in high-level decision-making for satellite deployment. This integration of supervised learning provides valuable insight for initiating large-scale constellations, specifically for the deployment stage just after orbit insertion.