In this work metabolic engineering strategies for maximizing L-(-)-carnitine production by Escherichia coli based on the Biochemical System Theory and the Indirect Optimization Method are presented. The model integrates the metabolic and the bioreactor levels using power-law formalism. Based on the S-system model, the Indirect Optimization Method was applied, leading to profiles of parameter values that are compatible with both the physiology of the cells and the bioreactor system operating conditions. This guarantees their viability and fitness and yields higher rates of L-(-)-carnitine production. Experimental results using a high cell density reactor were compared with optimized predictions from the Indirect Optimization Method. When two parameters (the dilution rate and the initial crotonobetaine concentration) were directly changed in the real experimental system to the prescribed optimum values, the system showed better performance in L-(-)-carnitine production (74% increase in production rate), in close agreement with the model's predictions. The model shows control points at macroscopic (reactor operation) and microscopic (molecular) levels where conversion and productivity can be increased. In accordance with the optimized solution, the next logical step to improve the L-(-)-carnitine production rate will involve metabolic engineering of the E. coli strain by overexpressing the carnitine transferase, CaiB, activity and the protein carrier, CaiT, responsible for substrate and product transport in and out of the cell. By this means it is predicted production may be enhanced by up to three times the original value.