Climate is related with vegetation on regional and even global scales, and climate affects spatial distribution of vegetation. Studies on correlation between NDVI and changes in climate using different time scales and different regional condition are helpful to reveal the effects of climate changes on vegetation. This study investigated the characteristics of vegetation change and its response to precipitation and temperature change during the 29 years (1981-2009) in the north of Tianshan Mountains in Xinjiang Province by the method of wavelet cross-correlation technique combined with trend analysis. The explored study area included Irtysh River, Bortala Valley, Ili River Valley, the northern slope of the Tianshan Mountains and the western Junggar Basin. These regions are far from the sea and surrounded on three sides by the Tianshan Mountains, Tarbagatai Mountains, and Altai Mountains, creating a special climate system which was an ideal environment for the study of regional climate-vegetation interactions. The daily meteorological data from January 1, 1981 to December 31, 2009 were collected from NOAA/AVHRR NDVI digital images (1981-2001) and SPOT-4 VEGETATION NDVI digital images (1998-2010). The results showed that in the northern Tianshan Mountains, temperature and precipitation change had similar tendencies except for the northern slopes of the Tianshan Mountains and in the Junggar Basin. The increase in the humid index decreased from the southwest to the northeast. Temperature increase may be primarily responsible for regional drying, while precipitation changes directly affected regional dryness and wetness. The annual mean NDVI distinctly differed among the six regions and it increased in the Bortala Valley, Ili River, Junggar Basin, and the mountains west of the Junggar Basin. The annual maximum NDVI decreased in all the regions except for the Junggar Basin. The correlation between early changes in climate and subsequent vegetation had multiple-time-scale characteristics, and all the regions showed similar overall correlations on different time scales. Time lag was another characteristic affecting the process above. The suitable time scale and time lag reflecting the response of vegetation to precipitation change was 15 ten-day periods and 5 ten-day periods except for the western Junggar Basin where time lag was 2 ten-day periods. The useful time scale and time lag reflecting the response of vegetation to temperature was 15 ten-day periods except the western Junggar Basin where the time scale was 10 ten-day periods and 2 ten-day periods, respectively. The correlation between NDVI change and early changes in temperature (0.587) was higher than that between NDVI and early changes in precipitation (0.456). The response of vegetation to climate change suggested that the effects of climate change on vegetation could differ greatly with time lags. This study could provide theoretical support for vegetation coverage monitoring, and enrich the knowledge about the complex climate-vegetation relationships.