Dynamic optimization is an important topic in chemically and biochemically reacting systems, notably prevalent in optimizing chemical/biochemical reactors for maximizing yield, minimizing cost, etc. The high-dimensional and complex nature of chemical processes makes it extremely difficult to ascertain analytically optimal solutions for such a wide range of problems. As such, machine learning-based approaches are employed to solve and obtain the optimal trajectory of control variables in the chemical process. Traditionally, discrete optimization techniques are employed for the task of solving ODE-based dynamic optimization problems, which work by converting a continuous range of values into a discrete set of points, through which a trajectory must be optimized. In this paper, we explore and evaluate the application of Deep Neural Networks (DNNs) in solving such chemical/biochemical problems. Combined with the Runge–Kutta 4th order method of ODE approximation, DNNs are trained to learn a smooth and continuous function representation of the control variable over time, which, for a given system of ODE constraints, optimizes the specified objective. We test our proposed methodology on four real-world-based case studies and compare our results with other literature employing traditional discrete optimization methods. A comparison of the results obtained reveals that the proposed DNN-based methodology yields results comparable to those obtained from state-of-the-art discrete methods. We further discuss some challenges and limitations faced in our methodology and potential ways to mitigate them, including the difficulty of DNNs in capturing non-smooth function profiles, and the handling of boundary constraints on variables in the chemical process. Finally, we raise a call for research into further evaluations and applications of neural networks in solving dynamic optimization problems, including the employment of hybrid models and advanced neural network architectures.