2020年是我国全面打赢脱贫攻坚战的收官之年。评估减贫效果是当前验收工作的重点，并对探索解决相对贫困的长效机制具有重要意义。本文通过生产精准扶贫阶段（2014-2020年NPP-VIIRS）夜间灯光遥感年合成数据，构建县域夜间灯光指数和夜间灯光变化指数，分别探讨了我国831个国家级贫困县和14个集中连片特困区的减贫效果。结果表明，2014年以来我国大部分贫困县的经济水平得到显著提高，减贫效果突出。然而，仍有108个贫困县夜间灯光强度呈现负增长趋势，主要位于西部地区的集中连片特困区交界处，西部地区内部出现南北发展不平衡现象。14个集中连片特困地区的夜间灯光亮度变化呈现出基数小增速快（I型）、基数大增速快（II型）、基数大增速慢（III型）和基数小增速慢（IV型）种类型，且在集中连片特困区交界处和省级行政边界交汇处呈现高高集聚和低低制约的空间分布格局，交界处的贫困县易被边缘化。进一步分析表明，实施基础设施扶贫、特色产业扶贫、资产收益扶贫（光伏扶贫）、易地搬迁扶贫这四类脱贫路径的贫困县夜间灯光变化明显。

Objective: By the end of 2020, China has scored a victory in its fight against poverty. Evaluate the poverty reduction effect is not only condutive to issue implementation plans for acceptance of poverty exit, but also of great imprtance to build a long-term and effectual mechanism to solve the relative poverty. Method: By synthesizing an annual dataset of NPP-VIIRS nighttime light (NTL) data from 2014 to 2020, we developed a county-level NTL index to investigate the poverty reduction effects of 831 national-level poverty-stricken counties and 14 concentrated contiguous poverty-stricken areas in China. Result: The results indicated that the economic level of most poverty-stricken counties in China has been improved significantly during the study period, and the poverty reduction effect is prominent. However, there are still 108 poverty-stricken counties with negative growth trend of NTL intensity, which are mainly located at the junction of concentrated and contiguous poverty-stricken areas in the western region. Besides, the NTL intensity development between the north and south parts of western regions is unbalanced. Four NTL development modes, i.e., small NTL base with a rapid growth rate (mode I), large NTL base with a rapid growth rate (mode II)), large NTL base with a slow growth rate (mode III), and small NTL base with a slow growth rate (mode IV) were identified in the 14 concentrated contiguous poverty-stricken areas. It was also interesting to find that the high concentration mode and low restriction mode were distributed at the junction areas among different provincial administrative boundaries. Besides, poor counties along the border are vulnerable to marginalization. Further analysis indicated that significant NTL changes were observed in the poverty-stricken counties with four poverty alleviation paths: infrastructure poverty alleviation, characteristic industry poverty alleviation, asset income poverty alleviation (photovoltaic poverty alleviation) , and relocation poverty alleviation.