P. putida emerged as one of the most promising production hosts for a wide range of chemicals, due to its fast growth with a low nutrient and cellular energy demand, considerable metabolic versatility, ability to grow in wide range of chemicals, suitability for genetic manipulations and its robustness and high flexibility to adapt and counteract different stresses. One of the main advantages of P. putida compared to commonly used industrial hosts such as E. coli is its superior tolerance to toxic compounds such as benzene, toluene, ethylbenzene, xylene and other hydrocarbons. In this work, we developed a large-scale kinetic model of P.putida to predict the metabolic phenotypes and design metabolic engineering interventions for the production of biochemicals. We first performed a gap-filling and thermodynamic curation of the genome-scale iJN1411 model of P. putida KT2440. The redGEM and lumpGEM algorithms for the systematic reduction of stoichiometric genome-scale models are then applied to the curated iJN1411 to derive a consistently reduced large-scale stoichiometric model of P. putida. Using this model as a scaffold, we next employed the ORACLE framework to generate a population of large-scale kinetic models around the experimentally observed steady state. To illustrate the predictive capabilities of these models, we performed two studies. First, for a wild-type strain of P. putida KT2440 growing on glucose under aerobic conditions, we computed metabolic responses to several single-gene knockouts, and the developed kinetic models successfully captured the experimentally observed phenotypes. In the second study, we proposed metabolic engineering interventions for improved robustness of this organism to stress conditions. Overall, the results from these studies suggest that the developed models of P. putida metabolism can successfully be used for metabolic engineering design.