Abstract: The cold chain logistics demand system of urban agricultural products in Beijing has many complex characteristics, such as non-linearity, few historical data, and many influencing factors, while support vector machine has outstanding advantages in solving the problems of small samples, non-linearity and high-dimensional pattern recognition. Therefore, the support vector machine model was introduced to train the data of the cold chain logistics demand of agricultural products in Beijing from 2000 to 2014, and then the cold chain logistics demand of agricultural products in Beijing from 2015 to 2020 was forecasted. The results showed that the support vector machine model can effectively fit the complex trend of agricultural products cold chain logistics demand system in Beijing, which could provide quantitative decision for agricultural products cold chain logistics planners and the government.

Key words: agricultural products, cold chain logistics needs, support vector machine, grey correlation analysis, prediction model