Permeability is a critical parameter in geotechnical and hydrological studies, representing a soil's capacity to transmit water through its pore spaces. The Hazen formula, an empirical method based on the D10 value from sieve analysis, is widely used to estimate the permeability coefficient. However, its accuracy is strongly influenced by grain shape and size distribution, which vary significantly across different geological settings such as volcanic, fluvial, and coastal environments. This study evaluates the applicability of the Hazen formula for sands derived from these three environments. The novelty of this research lies in its integrated approach, combining visual morphological analysis using sand grain imagery, particle size distribution tests, permeability estimation via the Hazen formula, and validation through constant head laboratory tests. Nine sand samples were analyzed: four from mountainous regions, two from rivers, and three from coastal areas. The findings reveal that sands with angular grains and wide gradation ranges exhibit considerable discrepancies between estimated and measured permeability values. Conversely, sands with rounded and well-sorted grains produce more consistent results. These outcomes suggest that the Hazen formula remains reliable for clean, uniformly graded sands with rounded to sub-rounded particles. This study highlights the need to consider grain morphology when applying empirical permeability estimation methods.