We develop an energy-based theoretical framework to predict jet morphologies in electrospinning. The charged polymer jet is modeled as a one-dimensional elastic curve subjected to a uniform electric field. The model includes the external-field energy, an effective local representation of curvature-dependent electrostatic self-interaction, and the bending and torsional elastic energies of the viscoelastic jet. To address the mathematical complexity arising from torsional effects and three-dimensional geometry, the resulting equations are solved using physics-informed neural networks (PINNs). The model identifies three distinct morphological regimes governed by the competition between electrostatic and mechanical energies: a straight configuration at low stiffness, a three-dimensional spiral at intermediate stiffness, and a planar spiral at high stiffness. The predicted spiral envelopes are consistent with the experimental trend and . This work provides a unified energetic interpretation of electrospinning jet morphologies and highlights the potential for integrating physics-based modeling with machine learning for complex electrohydrodynamic systems.