Abstract: Experiments on the influence of different weed populations on cotton yield were conducted in saline and alkaline land in Gaotang, Shandong Province, China in 2011 and 2012. There were four replicates of four treatments: weed free, grass weeds, broadleaf weeds and mixed weeds. The cotton varieties were Lumianyan 28 in 2011 and Lumianyan 37 in 2012. Cotton yield losses were 24.3% and 20.3% in grass weed plots (mainly Eleusine indica, Setaria viridis and Phragmites communis) at weed densities of 202.0 and 291.6 stem/m², and were 27.1% and 60.2% in broadleaf weed plots (mainly Chenopodium album, Portulaca oleracea and Eclipta prostrate) at densities of 230.5 and 122.2 stem/m² in 2011 and 2012 respectively. The yield loss was 37.5% in mixed weed plots (mainly C. album, E. indica and Echinochloa crusgalli) at 277.0 stem/m² in 2011, and was 68.4% (mainly C. album, P. oleracea and E. indica) at 201.0 stem/m² in 2012. The damage rate per stem caused by broadleaf weeds was higher than that caused by grass weeds. Weeds reduced the height of cotton plants, number of cotton bolls, boll branches and weight of bolls. The number of bolls per plant in weed free plots was 11.5 on average in both 2011 and 2012. The number decreased by 34.0% in 2011 and 72.0% in 2012 in mixed weed plots, 11.5% and 53.6% in broadleaf weed plots, 10.8% and 12.1% in grass weed plots. The height decreased by 0.4%-18.7% in weed plots compared with that in weed free plots. The number of boll branches decreased by 11.9%-18.0%, and the weight of bolls decreased by 6.3%-18.2%. In addition, weeds also reduced the survival rate of cotton seedlings by 0-3.6%. No effect was observed on ginning outturn and seed index. The influence of weed populations on cotton yield was closely related to weed species, density and biomass as well as to cotton variety and climatic factors.