This article present outstanding results given by the new application of Artificial Immune Systems in trend change detection in time series. Author?s system (GRASICA3), has been evaluated on a financial time series containing daily quotas of the main index of the Warsaw?s stock exchange (WIG) and additionally on a synthetic time series generated using the Monte Carlo method. Very good results which have been obtained (>60% of accuracy in trend change signals) are compared to results of other systems known from a bibliography, designed by Gutjahr and Kingdon. In the next stage the author provides a comparison of the GRASICA3 results to results given traditional statistical modeling methods such as, a very popular Box-Jenkins and Arima X-12 algorithms.