Industries are moving toward green operation and maintenance to help reduce their carbon footprint. This transition is associated with new challenges. One of these challenges is energy storage, where lithium-ion batteries are commonly used today. To ensure the safe and reliable use of lithium-ion batteries, the state-of-charge (SOC) is an important battery state to estimate accurately. In this work, a model capable of predicting battery voltage, SOC, and temperature during dynamic discharge conditions within a discharge cycle is developed. This is achieved by formulating the problem as a state-space representation and model the state transition function using a multiple-output Gaussian process regression (GPR) model. The system inputs are the instantaneous current along with the cumulative charge drawn from the battery since the beginning of the battery’s life. The model is capable of predicting the available SOC with an average root-mean-squared error (RMSE) of 0.039 for all the discharge cycles over the battery’s complete lifetime. The average RMSE for voltage predictions is 0.019. A Monte Carlo simulation with different initial values for the model is also performed to investigate how the uncertainty in initial values will propagate throughout the predictions. The model is able to converge to the predicted value with the correct initial values after 20 predictions for SOC and 15 predictions for voltage.