The adsorption of direct fast black onto acid-thermal modified sepiolite was investigated. Batch adsorption experiments were performed to evaluate the influences of experimental parameters such as initial dye concentration, initial solution pH and adsorbent dosage on the adsorption process. The three-factor and three-level Box-Behnken response surface methodology (RSM) was utilized for modeling and optimization of the adsorption conditions for direct fast black onto the acid-thermal modified sepiolite. The raw sepiolite was converted to acid-thermal modified sepiolite, and changes in the fourier transform infrared spectrum (FTIR) adsorption bands of the sample were noted at 3435 cm and 1427 cm . The zeolitic water disappeared and the purity of sepiolite was improved by acid-thermal modification. The decolorization rate of direct fast black adsorbed increased from 68.2% to 98.9% on acid-thermal modified sepiolite as the initial solution pH decreased from 10 to 2. When the adsorbent dosage reached to 2.5 g·L , 2.0 g·L , 1.5 g·L and 1.0 g·L , the decolorization rate was 90.3%, 86.7%, 61.0% and 29.8%, respectively. When initial dye concentration increased from 25 to 200 mg·L , the decolorization rate decreased from 91.9% to 60.0%. The RSM results showed that the interaction between adsorbent dosage and pH to be a significant factor. The optimum conditions were as follows: the adsorbent dosage 1.99 g·L , pH 4.22, and reaction time 5.2 h. Under these conditions, the decolorization rate was 95.1%. The three dimensional fluorescence spectra of direct fast black before and after treatment showed that the direct fast black was almost all adsorbed by the acid-thermal modified sepiolite.