An approach combining petrology and statistics (cluster analysis) is developed to better distinguish a series of subfacies in surficial and short core samples from Idku lagoon in the northwestern Nile delta, Egypt. The method integrates lithology, physical and biogenic structures, grain size, and composition of the sand-size fraction. Seven lithofacies, eight grain-size types and six clusters reveal comparable geographic distributions that are a direct response to geomorphic setting and depositional processes, both of which have been modified by man. Although positioned in an arid setting and affected by anthropogenic factors (land reclamation, pollutants), regional subfacies distribution patterns in Idku lagoon are largely a function of the spatial configuration of freshwater-dominated to brackish to higher salinity-dominated (near inlet) zones typical of lagoons receiving significant freshwater input. Within Idku, a northern depositional province near the inlet and along the coastal sand barrier is defined by associations of terrigenous and coarse-grained sediment types. A southern province, bounded by lagoon margins and marshes, is characterized by finer-grained sediment types, with a high biogenic content, recording lower energy levels. The petrological-statistical approach is used to identify and interpret lagoon subfacies in older (to > 6000 years B.P.) core sections in ldku region and in former Abu Qir lagoon. Additionally, cluster analysis serves to determine specific lagoon subfacies from hand samples alone, whether from surface grabs or cores; this is based on a comparison of lithofacies and clusters determined for surficial and short core samples in the ldku region. To test the method, a separate cluster analysis, also using surficial and short core samples, was used to define Manzala lagoon subfacies in the northeastern Nile delta. Results are generally comparable to those obtained in the ldku study, and show a close relation among Manzala lagoon subfacies, geomorphic configuration and depositional processes. The petrological-statistical method could be applied to other regions, and also to the geological record.