Energy consumption management shapes the route to meet greenhouse gas emission reduction goals. Although certain types of or even total energy consumption monitor have been realized using nighttime light remote sensing data, there still leaves a gap to model energy consumption in different sectors timely and broadly. Using parcel oriented temporal linear unmixing method (POTLUM), nighttime light was decomposed into light from different land uses as sources. A pilot study in Shanghai shows that in each year, statistical energy consumptions in living, secondary and tertiary sectors correlate well with unmixed nighttime light from 2014 to 2018 (R2=0.75, 0.61, 0.24, 0.99, 0.71, respectively). This study proves the capability of unmixed nighttime light to estimate sectoral energy consumption using POTLUM, which helps to better monitor energy consumption and serves GHG emission reduction.