Under the background of changing global environment and frequent anthropogenic activities, rainfall-runoff relationship has suffered tremendous change in many regions of China, particularly in arid and semi-arid regions. The Jinghe River basin (JRB) in the middle of the Loess Plateau, which is the major base of grain production in Shannxi Province, has experienced significant changes in hydro-climatic variables and the relationship between them during the past decades, causing many water resources problems. In this paper, we focused on analyzing the variation characteristics of rainfall-runoff relationship in JRB, and quantitatively assessing the effects of climate changes and human activities on the runoff reduction. It was of great importance for mastering the evolution of hydrological cycle, water resources planning and management, and water and soil conservation in JRB. The monthly rainfall and runoff data were used in the study and summed annually to investigate the variation characteristics in 1960-2010. To detect the change points of rainfall-runoff relationship, the sliding partial correlation coefficients method was proposed and the change points were further confirmed by the double mass curve method. On the basis of the studies, the relative changes of rainfall, runoff, and rainfall-runoff relationship were analyzed around the change points. Moreover, we also introduced the Archimedean Copula function, from the perspective of probability, to analyze the probability of synchronous-asynchronous encounter of rainfall and runoff in JRB. What caused the variation of rainfall-runoff relationship? Climate change or human activities? To figure out this, the method of slope change ratio of cumulative quantity (SCRCQ) was adopted in this paper. According to these results, we obtained the following conclusions: (i) Annual runoff and rainfall presented wavelike decrease change. Nevertheless, annual runoff declined more greatly than rainfall; especially, in some periods annual rainfall increased, however, runoff decreased. This illustrated that some non-rainfall factors caused the occurrence of phenomenon in JRB; (ii) Change points of rainfall-runoff relationship occurred in 1996 validated by the sliding partial correlation coefficients method and double mass curve method. Meanwhile, we could ascertain the conclusion by the evidence of the process of human activities in JRB, which presented the increasing industrial, agricultural and domestic water consumption and large-scale water conservancy and soil conservation projects since 1990s; (iii) To compare the magnitude of the changes of rainfall, runoff and rainfall-runoff relationship, the hydrological variables series were divided into the periods of 1960-1996 and 1997-2010. Compared to runoff and rainfall in 1960-1996, the annual mean runoff and rainfall in 1997-2010 dropped 43.47% and 6.64%, respectively. The change of monthly rainfall had no significance in different periods. On the contrary, monthly runoff decreased more greatly and the difference was smaller in each month i.e. the runoff distribution for each month was more uniform. The correlation between rainfall and runoff diminished and runoff yield decreased under the same rainfall. When precipitation was above 400 mm, the relative change rate of runoff was above 40% in 1997-2010 compared to that in 1960-1996; (iv) The joint distributions of rainfall and runoff constructed by the Archimedean Copula function were different in 1960-1996 and 1997-2010, and the distributions in the 2 periods were fitting Clayton and Gumbel-Hougaard Copula better, respectively.