The paper presents the approaches to the use of system analysis methods in modelling of hand-arm vibration syndrome (HAVS) in workers exposed to local vibration. The practice of using regression, information-entropy models artificial deep neural networks based on deep learning taking into account the dose of local vibration as an exposure factor is shown. The construction of HAVS occurrence model in the form of a multi-parameter regression approximated using a multilayer neural network is considered. A binary classifier that allows a set of attributes to attribute a person to a group of healthy people or to a group of people with HAVS is built. After training, the HAVS dynamics model made it possible to obtain a qualitative picture of the dependence of changes in the endocrine system on the individual experience dose of vibration and to determine the period of increased risk of pathology. The method applying to the occupational diseases differential diagnosis support system integration is discussed. Further study directions are also outlined.