Despite the recurrence of fingermark dating issues and the research conducted on fingermark composition and aging, no dating methodology has yet been developed and validated. In order to further evaluate the possibility of developing dating methodologies based on the fingermark composition, this research proposed an in-depth study of the aging of target lipid parameters found in fingermark residue and exposed to different in fluence factors. The selected analytical technique was gas chromatography coupled with mass spectrometry (GC/MS). The effects of donor, substrate and enhancement techniques on the selected parameters were firstly evaluated. These factors were called known factors, as their value could be obtained in real caseworks. Using principal component analysis (PCA) and univariate exponential regression, this study highlighted the fact that the effects of these factors were larger than the aging effects, thus preventing the observation of relevant aging patterns. From a fingermark dating perspective, the specific value of these known factors should thus be included in aging models newly built for each case. Then, the effects of deposition moment, pressure, temperature and lighting were also evaluated. These factors were called unknown factors, as their specific value would never be precisely obtained in caseworks. Aging models should thus be particularly robust to their effects and for this reason, different chemometric tools were tested: PCA, univariate exponentialregression and partial least square regression(PLSR). While the first two models allowed observing interesting aging patterns regardless of the value of the applied influence factors, PLSR gave poorer results, as large deviations were obtained. Finally, in order to evaluate the potential of such modelling in realistic situations, blind analyses were carried out on eight test fingermarks. The age of five of them was correctly estimated using soft independent modelling of class analogy analysis (SIMCA) based on PCA classes, univariate exponential linear regression and PLSR. Furthermore, a probabilistic approach using the calculation of likelihood ratios (LR) through the construction of a Bayesian network was also tested. While the age of all test fingermarks were correctly evaluated when the storage conditions were known, the results were not significant when these conditions were unknown. Thus, this model clearly highlighted the impact of storage conditions on correct age evaluation. This research showed that reproducible aging modelling could be obtained based on fingermark residue exposed to influence factors, as well as promising age estimations. However, the proposed models are still not applicable in practice. Further studies should be conducted concerning the impact of influence factors (in particular, storage conditions) in order to precisely evaluate in which conditions significant evaluations could be obtained. Furthermore, these models should be properly validated before any application in real caseworks could be envisaged.