The aim of this study was to use a high-density genetic map to identify QTL for panicle architecture-related traits in sorghum， which would enable marker-assisted selection and the exploration of the genetic mechanisms of key genes underlying these traits. 205 recombinant inbred lines （RILs） from a cross between BTx623， an American cultivar， and Hongyingzi， a cultivar for brewing liquor， were used to investigate five panicle traits in five environments （Guiyang， Anshun and Ledong） from 2020 to 2021. These traits include panicle length （PL）， panicle handle length （PHL）， number of cob nodes （NCN）， number of primary branches （NPB）， and length of the longest primary branch length （LLPB）. The inclusive composite interval mapping （ICIM） method was used to identify QTL， enabling the identification of 61 QTL at 45 unique loci. Of these， 14、10、8、11 and 18 QTL were related to PL， PHL， NCN， NPB， and LLPB， respectively. Nineteen important QTL were identified across multiple traits or environments and were distributed on chromosome 1（3）， 3 （4）， 4 （2）， 5 （1）， 6 （4）， 7 （1）， 8 （3）， and 9 （1）. In addition， 13 orthologous genes for rice panicle architecture were found in or near 12 of the 19 QTL， including DEP1， RGN1， OsPID， OsSPL7 and WTG1. These findings provided a basis for gene cloning and function verification in future.