Triglycerides are important constituents of food products showing polymorphic solid transitions leading to different conformational arrangements of molecules. Understanding the effects of macroscopic process conditions on conformational properties of triglycerides at the molecular level is of fundamental importance to improve industrial managements involving fats mixtures (one of the most important is that regarding the chocolate production process). In this thesis microscopic mathematical models are built to investigate the dependence of triglycerides conformations on environmental conditions. A detailed description of the microscopic dynamics of molecules is obtained through an atomistic model. This model is able to predict some physical properties of mixtures of triglycerides, such as density and viscosity. A coarse-grained model is developed to execute molecular dynamics simulations for longer time allowing the analyses of the conformational properties of molecules. The development of the coarse-grained model was tackled by means of statistical tools starting from the atomistic one. The coarse-grained model is able to highlight the dependence of triglycerides molecules conformations on environmental conditions which mimic typical food process conditions.