The Johnson‐Cook material model is a robust material model which has demonstrated its usefulness in describing material behaviour over large ranges of strains, strain rates and temperatures. During machining the material in the shear zone undergoes strains of more than 200%, strain‐rates of the order of 106 per second or more and a temperature rise of several hundreds of degrees Celsius. The determination of the Johnson‐Cook parameters, which are needed to describe the material behaviour in the severe conditions found during machining, has proved to be challenging, even using the state‐of‐the‐art experimental methods. Recent experimental methods rely on data obtained from strains of around 50% and strain rates of the order of 103 per second. In this paper, an inverse method for determining the Johnson‐Cook parameters from machining simulations is described. To demonstrate the concept, a finite element model of orthogonal cutting is created and a particular Johnson‐Cook parameter set is used for the simulation. It has been shown earlier that multiple Johnson‐Cook parameter sets exist which give rise to almost indistinguishable chips and cutting forces for a single set of cutting parameters. In order to eliminate some of these different sets, machining simulations are carried out for two different rake angles. Using the Levenberg‐Marquardt optimisation algorithm, the original Johnson‐Cook parameter set is re‐identified. In order to achieve this, the chip morphology and the cutting force are used to construct the objective function for minimisation. To determine the direction of the steepest descent, the Jacobian matrix is determined numerically with respect to the Johnson‐Cook parameters.