High energy planetary ball milling was applied to prepare sono-Fenton nanocatalyst from natural martite (NM). The NM samples were milled for 2-6h at the speed of 320rpm for production of various ball milled martite (BMM) samples. The catalytic performance of the BMMs was greater than the NM for treatment of Acid Blue 92 (AB92) in heterogeneous sono-Fenton-like process. The NM and the BMM samples were characterized by XRD, FT-IR, SEM, EDX and BET analyses. The particle size distribution of the 6h-milled martite (BMM) was in the range of 10-90nm, which had the highest surface area compared to the other samples. Then, the impact of main operational parameters was investigated on the process. Complete removal of the dye was obtained at the desired conditions including initial pH 7, 2.5g/L BMM dosage, 10mg/L AB92 concentration, and 150W ultrasonic power after 30min of treatment. The treatment process followed pseudo-first order kinetic. Environmentally-friendly modification of the NM, low leached iron amount and repeated application at milder pH were the significant benefits of the BMM. The GC-MS was successfully used to identify the generated intermediates. Eventually, an artificial neural network (ANN) was applied to predict the AB92 removal efficiency based upon the experimental data with a proper correlation coefficient (R=0.9836).