Long-term experiments are indispensable for the analysis of the long-term effects of various crop production methods and technologies. The long-term experiments have been set up in Martonvásár by Győrffy Béla more than 60 years ago and can be considered as part of the national heritage. Measurements made on plots of long-term experiments (LTE) are generally taken each year in the case of crop yield and other plant measurements. While the ANOVA procedure may be applied to repeated measurement data, under restrictive assumptions, other statistical procedures based on multivariate methods are more realistic. A more satisfactory alternative is to perform a mixed model analysis using REML (residual maximum likelihood) methodology. In the long-term crop rotation experiment, we analysed the multifactorial distinction beetween the wheat and maize crop sequences vs. continous cropping using discriminant analysis. The effect of various fertiliser treatments and the year on the yield of maize was studied in long-term maize monoculture and maize-wheat diculture experiments.