Hard turning is a machining operation of finishing performed on material with a hardness higher than 45 HRC and it has been demonstrated to be a profitable alternative to grinding process allowing, respect to this latter, better surface quality, improved process control reliability, and faster production rates. The high pressure and sliding velocity at the tool-chip interface, in addition to the deformation energy generated, lead to the formation of high temperature zones localized in the cutting region. These thermal loads enhance harmful aspects on both cutting tool and machined material. On the tool, the abrasive-diffusive wear mechanism is promoted. On the workpiece, a local surface heat treatment of quenching is observed, upgrading the formation of a thin layer of over-quenched martensite, called white layer, due to its appearance under microscopical observation, that is undesirable due to its brittleness and related cracks generation. In addition to this, hard turning is usually performed in dry condition further augmenting machining temperature and tool wear rate. Because of the extremely high hardness of the machined materials, the employment of expensive ultra-high hardness cutting tools is mandatory. Thus, to make the process valuable, it is essential to optimize the tool usage by avoiding catastrophic failure and being able to forecast the tool wear evolution during cutting time as a function of the applied process parameters. One of the possible approaches for predicting the wear behaviour is finite element (FE) simulations of the turning process. In this work, the wear of polycrystalline cubic boron nitride (PCBN) tools when turning AISI 52100 hardened steel is investigated. An abrasive-diffusive wear model has been developed and implemented in a FE engine, by means of a dedicated subroutine able to update the geometry of the worn tool as a function of the cutting parameters, with the intent of simulating the tool wear behaviour, in terms of crater and flank wear extension. For the calibration of the model, the tool wear measurements achieved from an experimental campaign of orthogonal cutting tests have been used. The influence of process parameters on the wear rate has been examined, underlining that the most affecting parameter on tool wear is the cutting speed: the higher the cutting speed the higher the tool wear. Furthermore, it has been observed that the tool breakage was due to an excessive depth of the crater wear that modified the initial negative rake angle to a positive one. This modification changed the stress state of the tool from compressive to tensile, not well tolerated by PCBN tools, bringing to catastrophic failure. The tool wear simulation results are in good agreement with the experimental ones, validating the usability of the proposed computational methodology for crater and flank wear prediction. Therefore, the application of this technique allows the optimization of cutting parameters from a preventive FE simulation analysis avoiding the need of exploiting new costly and time-consuming experimental tests.