Nickle (Ni), copper (Cu), zinc (Zn) and cadmium (Cd) in eight sediment samples from mangrove area, Red Sea, Egypt were fractionated using diethylenetriamine pentaacetic acid (DTPA) and a sequential extraction procedure (SEP). These sediment samples were collected horizontally from two locations (four samples from each site) based on different layer visual characterizations. Sediments were initially characterised using different analytical and statistical procedures. DTPA-extract table metal percent of the total metal content varies widely from 1.93 : 25.6% for Ni, from 3.86 : 18.2% for Cu from 0.6 to 4.1 for Zn and from 3.36 : 11.2% for Cd depending on soil characteristics. There no consistent trend of metal extracted by DTPA as a function of sediment depths although Cd_DTPA (%) was increased by increasing sediment depths in site 1 and in site 2. Moreover, metal extracted by DTPA showed no consistent correspondence with any single fraction of SEP. The metal extracted by DTPA showed reasonable correlation with the summation of exchangeable (F1) and carbonate (F2) fractions in both sites. So it may be reasonable to conclude that most ‘available' metal is (presumably) surface-bound on CaCO₃, rather than occluded within CaCO₃ or present as a mixed solid-solution (Ca_1-xM_xCO₃). Approximately 84%, 30%, 70% and 95% of the variability in PTPA-extractable-Ni, Cu, Zn and Cd (mg kg^-1) values for a whole dataset, respectively were explained by the variation in soil pH, (%) SOM, (%) CaCO₃ and total metals content. Cd_DTPA showed the highest prediction performance while Cu showed the worst case of the regression model in which Cu availability is more likely controlled by either the source of Cu or other conditions rather than sediment properties.