Identification of different types of turns during field operation of off-road vehicles is critical in the overall vehicle development as it is helpful in identifying & optimizing machine performance, correct duty cycle, fuel economy, stability analysis, accurate path planning, customer usage pattern & designing the critical components, etc. In this study, a machine learning (ML) based methodology has been developed to detect the off-road vehicle turns using vehicle & GPS parameters. Three most common types of off-road vehicles turn conditions e.g., Straight line, Bulb turn, and Three-Point turn have been considered. Different vehicle parameters (like latitude & longitude, compass bearing, yaw rate, vehicle speed, swash plate angle, engine speed, percent load at vehicle speed, raise lower front & PTO channels) generated during field test have been used here. These vehicle parameters are further processed, analysed and used in ML learning model building. Four ML models e.g., SVM, K-NN, Gaussian Naïve Byes and Random Forest have been used here. Experimental results show that the present ML based methodology can identify most common vehicle turns considered in this study with a good accuracy.