The thawing time and lake ice completely covered time can directly reflect the climate change in the region, and the duration of the lake ice can also reflect the local climate change. The traditional field monitoring of lake ice requires a lot of manpower and material resources, and the timeliness is poor, so it is not suitable for long-term ice condition monitoring. Since 1960s, remote sensing image has been used because of its wide range, fast speed, low cost and so on. More and more widely used in all kinds of research, the long-term monitoring of lake ice can also use remote sensing monitoring. Based on the Landsat series of remote sensing data, the Neural Network Algorithm was used to study the remote sensing image of Namco lake in the lake ice segmentation. The lake ice condition of Namco lake was monitored from 1988 to 2016.