Optimization of ultrasound-assisted extraction technique was performed in order to measure the concentration of polycyclic aromatic hydrocarbons in smoked rice by the high-performance liquid chromatography with fluorescence detection (HPLC/FLD) device using the response surface methodology (RSM). The experimental work was designed using the central composite design (CCD) in order to investigate the effect of the type and volume of extraction solvent, type of clean-up cartridge, the type, and volume of clean-up solvent on the response (i.e., BaP recovery percentage). The data obtained from 72 experiment runs were analyzed by analysis of variance (ANOVA) statistics. The results showed that the volume of extraction solvent, type of clean-up cartridge, and type and volume of clean-up solvent are significant in the extracted model. Desirability function was used to optimize the separation process. The optimal conditions were 24 ml acetonitrile extraction solvent, cartridge C18, and 0.856 ml mixture of dichloromethane/hexane solvent (50/50) for the clean-up. The BaP index was measured in smoked rice samples of Hashemi and local Domsiah using an optimized selective method. The contamination of the samples ranged from 0.1 to 1.1 μg⁻¹ BaP. Contamination in all samples except two ones was below the maximum permitted level of the EU (i.e., 1 μg l⁻¹).