This paper introduces into the research on trends in climate elements the multivariate statistical methods, namely principal component analysis (PCA) and cluster analysis, and demonstrates the benefits of their use. We also introduce the idea of normalization of the trends by their confidence interval, which facilitates the comparison of trends among different variables and allows their fair treatment by the multivariate methods. The study is performed for 11 climate elements at 21 stations in the Czech Republic. Temperatures are rising in winter, spring and summer, but decreasing in autumn. Consistent with these changes are trends in sunshine duration, cloudiness and daily temperature range. Correlation analysis and PCA confirm the mutual relationships among the trends in these variables in all seasons. The two methods also uncover other groups of variables whose trends vary similarly at individual stations, which, however, occur in several seasons only, not throughout the year. This implies that the trend analyses based on seasonal values should be preferred to analyses of annual trends. The cluster analysis of stations allows the quantification of the spatial consistency of trends. The groupings of stations are spatially incoherent and seasonally variable, which indicates that local peculiarities of the relatively complex terrain affect the station trends to a considerable extent.