Bioaerosol particles spread widely in the air， and high concentrations of bioaerosols pose a great threat to human health. To achieve early warning and prediction of atmospheric bioaerosol concentration， this paper uses fluorescence lidar as the detection tool. Based on the acquisition of bioaerosol concentration profiles， combined with relevant parameters of the atmospheric environment， particle swarm optimization （PSO） and genetic algorithm （GA） are used to optimize the support vector machine （SVM） to establish a bioaerosol concentration profile prediction model. Using temperature， humidity， PM2.5， PM10， CO₂， SO₂， NO₂， O₃， wind speed and other related parameter data as inputs， and bioaerosol concentration profile data as outputs for model training， the prediction model parameter configuration is determined. New atmospheric environment parameters are reintroduced， and the trained model is used to predict the bioaerosol concentration profile， which is compared with the bioaerosol concentration profile detected by fluorescence lidar. At the same time， different algorithms are analyzed to optimize the model''s predicted bioaerosol concentration and its relative error.