The objective of this work was to understand and optimize the flocculation of a marine alga Nannochloropsis oculata with two cationic salts, aluminum sulfate (AS), and ferric chloride (FC). Based on single-factor and response-surface-methodology experiments, second-order polynomial models were developed to examine the effect of initial algal biomass concentration (IABC), pH, and flocculant dose (FD) on final solid concentration of algae (SCA). The experimental and modeling results showed that SCA favored low pH, which however was undesirable to biomass recovery rate. There existed a positive stoichiometric relationship between FD and IABC; higher IABC required higher FD, and vice versa, for higher SCA. Optimum flocculation conditions were predicted at IABC of 1.7 g/l, pH 8.3, and FD of 383.5 mu M for AS, and IABC of 2.2 g/l, pH 7.9, and FD of 438.1 mu M for FC, under which the predicted maximum SCA were 32.98 and 30.10 g/l using AS and FC, respectively. The predictions were close to validation experimental results, indicating that the models can be used to guide and optimize the flocculation of N. oculata using AS and FC as the flocculants.