Thermal stability of batch processes is a major factor for the safe and efficient production of polymers and pharmaceutical chemicals. The prediction of thermal stability for such nonlinear, non steady-state processes is unreliable when using most stability criteria found in literature. A new stability criterion K is proposed. This is derived for complex reaction networks based on the divergence criterion. Lyapunov exponents are an alternative method to predict thermal runaway behaviour. Embedding thermal stability criteria within Model Predictive Control (MPC) frameworks results in advanced control systems capable of intensifying batch processes safely, hence resulting in shorter processing times. It is shown that embedding criterion K within MPC results in more efficient computational times than embedding Lyapunov exponents. Lyapunov exponents potentially can be applied to systems different from chemical exothermic batch reactors due to the general mathematical form. The effect of parametric uncertainty for process control is of utmost importance for industrial application. It is shown that the use of scenario-based MPC and worst case MPC, together with criterion K and Lyapunov exponents, results in a robust control scheme capable of intensifying batch processes whilst keeping them under control subject to parametric uncertainty. Both, scenario-based and worst case MPC with these two criteria resulted in safe control capable of intensifying batch processes. It is found that worst case MPC embedded with criterion K results in the most computationally efficient robust control scheme for the intensification of processes considered in this work.