Since the sixties there have been numerous works dealing with the classification of different types of amber based on spectroscopical data. Most of them were restricted to a visual inspection of the fingerprints or simple spectral comparisons. This thesis is mainly focused on gaining fingerprints by means of Py-GC-MS, FTIR and DTA-TG generated data followed by multivariate statistical treatments. As a kind of data pretreatment the window summation technique was used. Moreover characteristic details of each type of fossil resin were worked out, including size-exclusion chromatographical (SEC) and mass spectrometrical (MALDI-TOF) measurements. The evaluation of the data showed that multivariate statistical tools can be successfully used for a classification of fossil resins and pattern recognition. The results gained from different analytical data showed partly complementary results, because they were mainly focused on the polymer matrix or the volatile components. Classification results given in this work are often comparable with paleobotanical and paleochemical studies gained by other workers. A further approach based on artificial ageing of recent resins showed an annealing to the behaviour of certain fossil resins. Significant changes in the fingerprints were potentially useful for chemical dating trials.