Determination of electrolyte concentrations (mainly potassium) in vitreous humour has long been considered an important tool in human death investigations for the estimation of the post-mortem interval (PMI). On the basis of its well known potential in ion analysis, capillary zone electrophoresis (CZE) has recently been applied to achieve a rapid and simultaneous determination of inorganic ions in this extracellular fluid. In the present work, artificial neural networks (ANN) were applied for modelling of the relationship of multicomponent CZE analysis of K+, NH4+, Na+, and Ba2+ ions in vitreous humour with PMI. In a study based on 61 cases with different causes of death and a known PMI ranging from 3 to 144 h, the use of ANNs considering all inorganic ion data from the human vitreous humour, achieved a substantial improvement of post-mortem interval prediction. Good linear correlation was observed (r2 = 0.98) and in comparison to the traditional linear least squares (LLS) method applied only to K+ levels in the vitreous humour, the prediction of PMI with ANN was improved by a factor of 5 from approximately +/- 15 h to less than 3 h.