This research evaluated the effectiveness and feasibility of two-phase (separation acid and methanogenic phases) anaerobic digestion of simulated sewage sludge using a UASB-UASB (upflow anaerobic sludge blanket) process. Predictive models of chemical oxygen demand (COD) (soluble) removal efficiency and CH4 gas production during loading maximization and the recovery (after failure) period were made. The optimum hydraulic retention time (HRT) for the A-UASB was about 1.0 d; for the M-UASB system, two different optimum HRTs were evident, depending on the operational mode. When the UASB-UASB system was run under the maximum loading possible, the best HRT in the M-UASB was about 2.0 d, to achieve high COD removal and concurrent optimum CH4 production; after deliberate overloading, to induce failure, and subsequent system recovery, the predicted optimum HRT in the methane unit was about 2.7 d, to achieve concurrent adequate CH4 production and COD removal. There also appeared to be a restructuring of the bacterial community inside the M-UASB, during the recovery period. For overall design purposes, optimum operating HRTs of 1 and 2 d, with an internal recycle rate of 1.6 and 2.5 times the influent flow rate, are recommended for A- and M-UASBs, respectively; a conservative organic loading rate of 19 kg COD (total)/(m3∙d) is suggested. Finally, the optimum HRTA-UASB/HRTM-UASB ratio was found to be about 0.63 and HRTA-UASB/HRTsystem was 0.38, which are within the boundaries of this research project. Key words: anaerobic sludge treatment, design criteria, optimum system loading capacity, two-phase anaerobic stabilization, upflow anaerobic sludge blanket.