Abstract: In order to solve the problems of measurement error and low accuracy of current carbon sink measurement methods, a carbon sink measurement method based on BP neural network is proposed. Firstly,the carbon sink data was preprocessed based on the carbon peak target to obtain the high accurate carbon sink data. The BP neural network model of carbon sequestration measurement was constructed to output the carbon sequestration data set according to BP neural network. Then, considering the carbon pool selection index, the carbon pool matching the carbon sink data set was selected. Finally, the goal of carbon sink measurement under the goal of carbon peak was achieved. Experiments showed that this measurement method had high fitting degree.The calculated value was closer to the original value. The relative error was between 1.01% and 1.71%, which was always below 2%. It proves that the accuracy of carbon sink measurement results of this method was high.