This work addresses the dynamic optimization of the production period of thermal crackers with respect to coke formation in the cracking coil and transfer line exchanger. Optimal time-dependent trajectories to feed rate, steam to hydrocarbon ratio, and reaction severity are calculated The net earnings based on the price of hydrocarbons, fuel, steam, decoking, and maintenance cost are maximized. All important operational constraints are included and the optimization problem is solved using parametrized free variable trajectories (piece-wise constant) and a standard SQP package. Rigorous distributed physical models are used and calculation show that dynamic optimization gives lip to 2% higher net earnings (earnings-expenses) than conventional steady-state optimization performed on the same models. This is in the same range as earnings reported from steady-state optimization implementations alone.