Abstract: Selecting the panel data of 13 prefectures and cities in Xinjiang from 2003 to 2018 as samples， based on the DEAmalmquist indexArc GIS visualization analysis， comprehensive analysis and comparison of Xinjiangs agricultural total factor productivity was made from time to vertical and geographical space. The results show： 1） The agricultural TFP fluctuates significantly and changes synchronously with TE； 2） The EF， PE and SE values have the same trend of change， and the overall lack of efficiency； 3） Only TE is efficient in all regions， and there are large differences in TFP between regions； 4） As time progresses， the agricultural TFP of various states in Xinjiang has declined， and the level of agricultural productivity has declined in recent years； 5） The four types of agricultural efficiency in various states have been identified， and reasonable agricultural development opinions have been put forward according to local conditions.

Key words: agricultural total factor productivity, temporal and spatial evolution, problem identification, DEAmalmquist index, Xinjiang

摘要： 选取新疆13个地州市2003—2018年面板数据为样本，基于DEAmalmquist指数ArcGIS可视化分析，从时间纵向和地域空间综合分析比较新疆农业全要素生产率。结果表明：农业TFP波动幅度明显，与TE同步变化；EF、PE和SE值三者变化趋势一致，整体缺乏效率；各地州仅TE有效率，地域间TFP差异大；随着时间推进，新疆各地州农业TFP下滑，近年农业生产力水平下降；识别出各地州农业效率的4种类型。据此，提出合理农业发展建议。

关键词: 农业全要素生产率, 时空演化, 问题识别, DEAmalmquist指数, 新疆