Atmospheric aerosols contribute to pollution and the deterioration of air quality. Among them, desert aerosols represent a major challenge, as they are capable of being transported over very long distances, and their effects on health, the environment and the climate system are well known. It is therefore important to model these aerosols correctly, as an accurate description of them is a major issue in many fields, including energy (solar radiation) and military defense, for example. This is the role of Transport Chemistry models such as MOCAGE (MOdèle de Chimie Atmosphérique À Grande Échelle) at Météo-France, which is capable of simulating chemical composition and atmospheric aerosols based on meteorological forcing and emission inventories. However, the prediction of aerosol concentration by these models still depends on uncertainties in meteorological forcing, atmospheric processes and emission sources. Infrared satellite observations from polar-orbiting instruments such as IASI have spectral ranges that are sensitive to desert aerosols and could improve their representation and monitoring in transport chemistry models. Previous work enabled us to carry out a study of IASI's sensitivity to desert aerosols using the RTTOV radiative transfer model, as well as the first experiments involving the direct assimilation of IASI radiances into MOCAGE for aerosols. In addition, the LISA laboratory has developed an algorithm, "AEROIASI", which provides the vertical profile of the extinction coefficient and the AOD of desert aerosols for each IASI observation in the absence of clouds. The aim of this study is to define the appropriate methodology for the direct assimilation of IASI spectral luminances into MOCAGE for the representation of desert aerosols. In this poster, we will present firstly how we used the methodology of LISA's AEROIASI product to define the spectral bands to be used, to assess any limitations of the radiative transfer model used in MOCAGE (RTTOV) and to select the relevant scenes to be used. Then, the parameters for assimilating IASI spectral radiances will be presented: draft and observation error matrices, assessment of the possible need for bias correction on observations, etc. All these preliminary assessments will be used to carry out data assimilation experiments on selected IASI observations, in order to improve the concentration fields of desert aerosols divided into 6 bins (aerosol sizes) in MOCAGE. Statistical evaluations will be carried out by comparing our results with other instruments or products. Results and evaluations of series of assimilation trials will be presented, as well as a way forward.