The invention discloses a low-carbon comprehensive evaluation method based on a neural network in a cross-border power trade environment. The method comprises the following steps: firstly, taking unitelectric quantity carbon emission generated in three power generation forms of coal power generation, wind power generation and photovoltaic power generation in a current region as an initial low-carbon evaluation index; by combining economic and new energy consumption indexes, establishing a discrete Hopfield low-carbon economic index evaluation model through a discrete Hopfield neural network analysis method, and evaluating the low-carbon benefit level of comprehensive new energy consumption of the region. According to the method provided by the invention, energy conservation and emission reduction evaluation of carbon emission generated in different energy power generation links can be realized, and on the basis of the evaluation result, comprehensive benefit evaluation is carried outon the power generation form of each region of the cross-border power trade environment in combination with economic and new energy consumption indexes.