摘要： In this work, Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) optimization techniques are used to retrieve water column inherent optical properties (IOPs), bottom reflectance and geometric depth from a simulated hyperspectral dataset. The synthetic dataset (N=6000) is created using two different bottom types (Sand and seagrass), three levels of depth, 20 concentrations of Chlorophyll – a (Chl-a) with a wide range of modelled IOP variations resembling natural clear and optically complex waters. Results indicate that PSO and GA showed similar retrieval results in IOPs. Bottom reflectance values from PSO are near to input bottom substrate values compared to GA. In case of Geometric depth also, average depth value retrieved by PSO are better than GA retrieved. In terms of processing time, PSO exhibited higher computational efficiency than GA by five times. Further studies can use PSO to process Hyperspectral remote sensing Imagery in shallow waters.