Vertical aerosol distributions remain poorly constrained in atmospheric chemistry models because most surface and satellite observations provide only near-surface or column-integrated information. We developed a three-dimensional variational data assimilation algorithm to use vertically resolved laser measurements from the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument to improve aerosol simulations. The system directly assimilates 532 nanometre total attenuated backscatter measurements into the Weather Research and Forecasting model coupled with Chemistry, using the four-bin Model for Simulating Aerosol Interactions and Chemistry aerosol scheme. We evaluated the system using two nighttime satellite observations over northern China in December 2019, representing moderate and heavy pollution conditions. Assimilation improved the simulated vertical distribution of aerosols along the satellite tracks. Compared with a simulation without assimilation, the aerosol optical depth correlation coefficient increased from 0.51 to 0.58, the root-mean-square error decreased from 0.28 to 0.24, and the mean bias improved from −0.11 to 0.01. For the heavy-pollution case, the initial mean surface concentration of fine particles increased from 85 to 107 micrograms per cubic metre, compared with an observed value of 151 micrograms per cubic metre. The improvement in surface fine-particle forecasts lasted approximately 14 hours. These results demonstrate the potential of satellite lidar measurements to improve three-dimensional aerosol analyses, while highlighting the need for broader evaluation because of limited satellite sampling, cloud effects, and observation uncertainty.