Soil‐working tools in agriculture are made of boron containing steels with high wear resistance and hardenability. Nevertheless, these tools are subject to high impacts, abrasive wear, fatigue and therefore prone to failure. To combine varying levels of properties within one component in as‐quenched condition can be beneficial for such products. To obtain this property variation a component must undergo a complex and controllable cooling. Therefore, the aim of this work is to obtain a microstructure gradient along two 15 mm thick steel plates in a newly developed test rig by Water Jet Impingement Technique to confirm its controllability and flexibility. Furthermore, a quenching simulation model was created for hardness prediction using phase transformation data from a machine learning tool. Microstructure variation was observed using light optical microscopy and the electron backscatter diffraction technique. Mechanical properties were studied through tensile tests and hardness measurements, and also compared with simulation results. 0.27 mass‐% C steel sample obtained almost fully martensitic state transitioning to a softer ferritic/bainitic condition, while the 0.38 mass‐% C steel sample resulted predominantly in a fully hardened martensitic state slightly showing ferritic and bainitic features along the sample. The quenching simulation model showed promising hardness prediction for both steels.This article is protected by copyright. All rights reserved.