Abstract: With intensive glyphosate application, its residues and consequent risks of soil health and ecological environment safety have received greater attention. The degradation kinetics of glyphosate in red soil aggregates with different sizes, as well as the interaction between physical and chemical properties of soil aggregates and the degradation of glyphosate, have rarely been studied. Thus, in this study, the degradation characteristics of glyphosate in red soil aggregates with different sizes were observed under laboratory conditions by particle pre-sieving, incubation in a controlled climatic chamber, and residue analysis via liquid chromatography-tandem mass spectrometry. The physical and chemical properties of the soil aggregates, such as contents of organic matter, total phosphorus, and available phosphorus, were tested according to the national approved methods and standards. The relationships between the physical and chemical properties of the aggregates and the degradation of glyphosate were further analyzed and compared in the same observation day. The results showed that 1) the glyphosate content decreased in the different aggregate particles during the observation period, following the single first-order kinetic degradation model. However, no significant differences were observed among different aggregate sizes. The half-life time of glyphosate in the different red soil aggregates ranged from 15.8 to 20.6 d, with a longer half-life time in the smallest aggregates (< 0.25 mm, 20.6 d). The aminomethylphosphonic acid (AMPA) content, the main metabolite of glyphosate, increased immediately and peaked on the 5^th day after glyphosate application, but no differences were found among different aggregates. However, the AMPA content changed and declined significantly in different aggregates after the 5^th observation day (P < 0.05). The contents of organic matter, total nitrogen, total phosphorus, and available phosphorus in different aggregates varied greatly, especially the available phosphorus content, which decreased with glyphosate degradation. 2) Correlation analysis and principal component analysis of glyphosate, aggregate size, and their properties showed that the residual glyphosate was significantly positively correlated with the content of available phosphorus (P < 0.05), and the AMPA content was significantly positively correlated with the activities of acid phosphatase and N-acetylamino-β-glucosidase (P < 0.05). There were no significant relationships between the aggregate size and the residuals of glyphosate, but a significant positive correlation was observed between the aggregate size and the AMPA content (P < 0.05). Furthermore, during the whole period of glyphosate degradation, the organic matter content, acid phosphatase, N-acetylamino-β-glucosidase, and β-glucosidase showed a significant negative relationship with the soil aggregate size (P < 0.05). In conclusion, the characteristics of the red soil aggregates affect the degradation kinetics of glyphosate, as well as the persistence of AMPA, especially the residuals in the smallest aggregates (< 0.25 mm). The contents of glyphosate and AMPA in the red soil aggregates were still high after 30 days, which may affect soil health. Glyphosate degradation was also closely related to phosphorus in the soil. Therefore, the fate of glyphosate under conditions of phosphorus deficiency or abundant soil should be explored to provide detailed information on glyphosate risk assessment in red soil.