This paper proposes a preconcentration procedure for determination of nickel in saline aqueous waste samples by flame atomic absorption spectrometry (FAAS). It is based on cloud-point extraction of nickel(II) ions as 2-(5-bromo-2-pyridylazo)-5-diethilaminophenol (Br-PADAP) complexes using octylphenoxypolyethoxyethanol (Triton X-114) as surfactant. The optimisation step was performed using a four-variable Doehlert design, involving the factors centrifugation time (CT) of system after addition of surfactant, solution pH, methanol volume (MV) added at micellar phase, and buffer concentration (BC). The analytical response used was absorbance, after volume correction. Using the established experimental conditions in the optimisation step the procedure enables nickel determination with a detection limit (3 delta/ S) of 0.2 microg L(-1), quantification limit (10 delta/ S) of 0.7 microg L(-1), and precision, calculated as relative standard deviation ( RSD) of 4.7 ( n=8) and 3.5% ( n=8) for nickel concentration of 1 and 5 microg L(-1), respectively. The preconcentration factor, determined from the ratio of the slopes of the analytical curves with and without preconcentration, is 74. The recovery achieved for nickel determination in the presence of several cations demonstrated that this procedure could be applied for analysis of water samples. The robustness was checked by using saturated fractional factorial designs, centred on the established experimental conditions in the optimisation step. The results of these tests demonstrated that the variables centrifugation time and buffer concentration are robust for modification by 10% and that solution pH and methanol volume are robust for 5%. Accuracy was evaluated by using the certified material reference SLEW-3 estuarine water for trace metals. The procedure was used for determination of nickel in saline effluents from oil refinery samples. Recovery results (95-104%) indicate that the procedure has satisfactory accuracy for nickel determination in these samples.