This paper measures the international trade efficiency of developing countries based on the data envelopment analysis (DEA) model, and explores the impact of digital transformation on trade efficiency differentiation using regression analysis. Relevant data of 19 developing countries, including China, are selected, and the trade efficiency at each stage is calculated separately using the three-stage DEA model in this paper. The regression model is constructed to quantitatively analyze the impact of digital transformation in the differentiation of trade efficiency of developing countries. From 2011 to 2020, the trade efficiency of each developing country shows a wave-like upward trend, and the average value of the comprehensive average efficiency in the third stage is 0.728, but only China, Peru and Colombia have a higher than average level of trade efficiency, which intuitively demonstrates the trade efficiency differentiation of developing countries. Differentiation. The overall regression results show that the elasticity coefficient of digital transformation on the international trade efficiency gap is -0.274, indicating that digital transformation has a greater effect on narrowing the trade efficiency gap than widening it. And in the subregional regression, the elasticity coefficient of digital transformation in Asia is 1.398, and the elasticity coefficients in Africa and Latin America regions are -0.953 and -0.603 respectively, and the digital transformation has significantly different impacts on trade efficiency differentiation in different regions.