Abstract: To investigate the impact of climate change on the climate suitability for soybean cultivation in Northeast China, daily meteorological data from 1961 to 2023, county-level soybean yield data from 1993 to 2022 and LightGBM model are used to study changes in climate conditions, drought and low temperature, as well as their comprehensive effects on the climate suitability for soybean cultivation. As a result of global warming, heat resources in Northeast China have significantly improved, and characteristics of disasters also change obviously. The average temperature during the growing period of soybeans in Northeast China shows a significant increasing trend, with an average temperature rise rate of 0.25 ℃·(10 a)^-1 from 1961 to 2023. The accumulated temperature reached 2884.9 ℃·d from 2021 to 2023, with an increase of 8.8% compared to the 1970s. The fluctuation in precipitation during the growth period of soybeans in Northeast China has increased, while the number of sunshine hours has decreased. The number of low-temperature days during soybean flowering shows a significant decreasing trend, with a reduction of 45.4% during 2014-2023 compared to the 1960s. There is no increasing trend in the number of drought days during the soybean growing period in Northeast China. LightGBM model is trained and utilized to identify climate suitability parameters for soybean cultivation in Northeast China. Soybean yields exhibit a negative correlation with the number of drought days during the growing period and the number of low-temperature days during the flowering stage. The correlation coefficient between soybean yield and the number of low-temperature days during the growing period of soybean is -0.23 (passing the test of 0.01 level). The yield of soybeans is positively correlated with accumulated temperatures (no less than 10 ℃) and the monthly average temperatures from May to September. In order to train the model, samples are divided into training and testing sets, with 80% of data allocated to the training dataset and the remaining 20% designated as the testing dataset. The correlation coefficient between the predicted index and the actual index in the model training set is 0.94 (passing the test of 0.01 level). The correlation coefficient between the predicted index and the actual index in the test set is 0.80 (passing the test of 0.01 level). This study utilized annual climate data from 1961 to 1990, 1991 to 2020, and 2014 to 2023 to drive the LightGBM model and calculate the changes in climate suitability indices for soybean cultivation. It demonstrates that, as a result of climate warming, areas suitable for soybean cultivation in Northeast China have significantly expanded to the west and the north. Suitable soybean planting sites comprised 90% of the total research sites from 1961 to 1990, and this figure increased to 95% of sites from 1991 to 2020.