Long-term experiments (LTEs) are early indicators of the sustainability of cropping systems and the effect of climate change. The LTEs set up in Martonvasar (HU) are now over 50 years. The effect of various crop production factors (crop rotation, fertilisation, genotype, sowing date and plant density) on the yield and yield stability of maize in different years was investigated, based on data series collected over several decades. The clarification of year effect is particularly important in relation to the possible effect of climate change. Our analysis approach was the mixed model ANOVA that incorporate the estimated serial correlation structure. Yield stability of treatments was examined by the regression method and by the AMMI model. The yield increasing effect of crop rotation was inversely proportional to the ratio of maize in the sequence and was greatest in the Norfolk type rotation, followed by alfalfa-maize-wheat, the maize-wheat and maize-alfalfa rotation. Yield and yield stability was higher in crop rotation comparing with monoculture. In dry years the yield decreased significantly, by around 30%, and changes were observed in the pattern of yield responses. In dry years lower rates of fertilisation were more efficient, and farmyard manure (FYM) alone and in combination with mineral fertiliser enhanced yield and yield stability. There were considerable fluctuations over the years in the N-fertiliser response of maize hybrids. Ex ante data that form the basis of farming decisions can only be obtained from LTEs. Sowing later than the optimum date led to a reduction in N fertiliser efficiency, which was more severe in dry years. Crop production factors contributed to the increase in maize yield in the following ratios (%): fertilisation 30.5, variety 32.6, plant density 20.2, weed control 14.2, soil cultivation 2.4. A non-decreasing trend in yield is necessary to call a system sustainable.