The feasibility of a renewable source for synthesis of mesoporous activated carbon (AC) as catalyst support in Fischer-Tropsch synthesis (FTS) has been investigated. For this purpose, machine learning was used to optimize the process conditions for synthesis of AC (through chemical activation) from biomass. Three different regression methods were carried out to optimize the activation time, temperature, and impregnation ratio. Random forest regression (RFR) resulted in the highest prediction accuracy with R2 of 0.88 and 0.97 for total pore volume (cm3/g) and mesoporosity (%), respectively. The prepared AC at optimum activation conditions (obtained by RFR) led to 90% mesoporosity. Prior to FT reaction, the mineral impurities of the AC were decreased using alkaline treatment. The performances of the Fe catalysts supported on AC (10Fe/AC and 20Fe/AC) were tested for FTS at 300 °C, 2 MPa, and gas hourly space velocity (GHSV) of 2000 h−1. The 20Fe/AC catalyst achieved 46.7% CO conversion and C5+ selectivity of 72.5%, indicating the promising potential of the biomass-based catalyst support with optimized textural features and modified surface for FTS. The 20Fe/AC catalyst showed superior FT activity and C5+ selectivity compared to the 20Fe/Al2O3, and Fe catalyst supported on commercial AC. For 20Fe/AC catalyst, liquid hydrocarbons in the range of C5–C50 were detected.