Abstract: Single particle analysis of environmental samples is an important analysis content in nuclear safeguards. Its purpose is to detect undeclared nuclear activities by analyzing the isotopic abundance of uranium bearing particles released into the environment during the nuclear fuel cycle. Since the nuclear activities characterized by the abundance of each single particle may be different, in order to obtain sufficient information, it is necessary to conduct isotopic analysis on as many uranium containing particles as possible in the samples, which requires that as many uranium bearing particles as possible be recovered to the sample pad during sample preparation, and in order to prevent the generation of mixed particles, it is required to have good dispersion between particles after sample preparation. As the first step of the whole environmental sample analysis process, the efficiency of the recovery method directly affects the efficiency of the whole particle analysis and the ability to detect undeclared nuclear activities. In this work, a sample preparation device for secondary ion mass spectrometry (SIMS) particle analysis was developed. Its basic principle is to recover micron particles to the sample pad through vacuum collision, and attach a layer of adsorption material to the sample pad to adsorb particles, so as to reduce the rebound phenomenon after particles hitting the sample pad. Based on the aerodynamic theoretical calculation, the pneumatic simulation software Fluent 15 0 was used to simulate the trajectory of particles in the recovery device. Through this method, the structure of the recovery device was optimized and the parameters of the recovery device were determined, so as to establish a new scanning recovery device with stable recovery, good particle dispersion and concentrated recovery area. On this basis, the effects of additional adsorption materials on the recovery and possible background interference were studied. The results show that for 1 μm uranium containing particles, the particle recovery is about 70% without considering the residual uranium particles on cotton cloth, this result is similar to that calculated by the simulation of aerodynamic software, because the software only considers the trajectory of particles after entering the device from the air inlet of the device, and does not consider the influence of residual particles on the recovery. When considering the residual uranium particles on cotton cloth, the particle recovery is about 45%. Using Apizon as adsorption material can effectively improve the recovery, while higher baking time and temperature and lower Apizon dropping amount can effectively reduce the background interference caused by polyatomic ions in SIMS analysis. At this time, its impact on particle main isotope analysis can be ignored.