In line with significant rise over the last decade in the body of data available from high-fidelity simulations and experiments, there has been an increasing interest across fields of science and engineering to employ data-driven methods and machine learning to develop interpretable and generalizable reduced-order models (ROMs) that can aid in explaining the observed complex, multidimensional phenomena and/or to enable prediction and forecasting of the systems’ behavior. Despite major advances in data-driven (DD)/machine-learning (ML) algorithms for physics modeling and dynamics discovery in the past years and the promising applications demonstrated across multiple scientific domains, particularly in fluid dynamics, the data-driven methods have not yet found a rigorous, widespread application for plasma physics, especially within the low-temperature technological plasmas communities. In this article, we aim to demonstrate the potential of two highly promising data-driven algorithms for plasma state forecasting and to report on the other research directions being pursued at Imperial Plasma Propulsion Laboratory (IPPL) on the broader subject of machine-learning-enabled/enhanced plasma modeling. First, we present results from linear-time-dynamics ROMs obtained from the Optimized Dynamic Mode Decomposition (OPT-DMD) method toward predicting plasma properties’ evolution in two test cases, one representing a radial-azimuthal cross-field discharge similar to that of a Hall thruster, and the other one corresponding to a Penning discharge configuration. Second, we will introduce a novel in-house developed approach, named the “Phi Method”, and demonstrate its capability for simultaneous forecasting of a coupled system of plasma state variables in a 1D azimuthal problem. Third and last, we provide an overview of the ongoing efforts on the “Sparse Identification of Nonlinear Dynamics” (SINDy) algorithm to develop reduced-order PDE/ODEs for plasma systems as well as on the use of super-resolution/subgrid-modeling techniques to enhance kinetic particle-in-cell simulations. We will conclude by highlighting how these DD/ML elements will underpin the laboratory’s overarching goal of establishing digital twin frameworks for spacecraft plasma propulsion systems.