【Aim】 Aedes albopictus is a highly invasive mosquito and the main medium of dengue contagion in Guangdong Province, South China; its potential distribution in this region can provide scientific evidence for the establishment of epidemic prevention strategies. 【Method】 Traditional methods do not take into account the relative importance of environmental factors. To address this problem, four combined models for predicting potential distribution were constructed: combinations of the four factor-weighting models (geographic detector, multivariable linear regression, principal component analysis, and factor analysis) and the analogy deviation model. The weights of environmental factors screened by correlation analysis were first divided by the four factor-weighting models and then combined with the analogy deviation value formula. Finally, the potential distribution areas of A. albopictus in Guangdong Province in southern China were predicted using GIS technology. 【Result】 The accuracy verification and prediction of the four combined models showed that the accuracy of the combination of the geographic detector model and the analogy deviation model (GDM-ADM) was the highest, with a mean AUC of 0.944 and a standard deviation of 0.008. Meanwhile, the GDM-ADM model prediction showed that the areas with a lower risk of invasion of A. albopictus were in the northern part of Guangdong Province, accounting for 4.05% of the total area, and most areas were under medium or medium-high risk, accounting for more than 85% of the total area, while Guangzhou, Foshan, and Dongguan in central Guangdong Province were at high risk, accounting for 8.77% of the total area. 【Conclusion】 Compared with the analogy deviation model, which does not consider the weights of the factors, the combined models that consider factor weights can effectively improve the accuracy of the prediction of potentially suitable areas for mosquitoes. The geographic detector model divides the weights of the factors by exploring spatial heterogeneity, which is more effective than traditional statistical models, and its combined model has the highest accuracy for the prediction of the potential distribution areas of A. albopictus..