Climate extremes have more direct and significant impact than average state on social and ecological systems. Using data from 68 meteorological stations in the arid region of northwest China (ARNC) for the past 50 years (1961–2010), we conducted Mann–Kendal nonparametric trend analysis on the time series of temperature and precipitation extremes in different regions of the ARNC. The analysis found that in the past 50 years, 1) overall for the ARNC, three temperature indices, including the annual mean temperature (Tav), annual average daily minimum temperature (Tnav), and annual average daily maximum temperature (Txav) all had increasing trends; 2) overall for the ARNC, three precipitation indices, including the annual mean precipitation (Pav), number of days for daily precipitation ≥ 10 mm (Pn10mm) and annual maximum number of consecutive wet days (Pxcwd) also all had increasing trends; 3) regionally, Tnav and Pn10mm in north Xinjiang considerably increased; in both north and south Xinjiang, annual maximum number of consecutive dry days (Pxcdd) considerably decreased, whereas Pxcwd considerably increased; and the main pattern of Hexi Corridor is that Txav considerably increased; 4) it appears that the increase of Tav in the ARNC in the past 50 years is related to the increase of Tnav and Txav; and 5) the increase of precipitation in north and south Xinjiang is a result of the joint effect of the increases of Pn10mm and Pxcwd. Generally, in ARNC, during the past 50 years the dynamics of climate extremes are closely related to the dynamics of climate average state, and are major contributors to the overall climate change.