The new approach for producing useful chemicals (i.e. methanol, acetone, acetic acid, phenol and m-cresol) from palm oil mill effluent has been proposed in this study. The liquid waste before entering a decantor was used as raw material. It was firstly separated to a solid cake and liquid. The obtained solid cake was dried and pyrolysed to produce bio-oil. The pyrolysed oil was subsequently reacted with the filtrated liquid to recover the target useful chemicals. Since the bio-oil had two phases (aqueous and oil), the reaction tests with a zirconia-iron oxide were performed separately to determine the optimal condition for each phase. The liquid waste before entering the decantor composed of water, oil and suspended solid. It had several large molecular hydrocarbons such as cellulose, hemicelluloses, lignin and fatty acids. From the filtration, the solid cake (18.25 wt%) and liquid (81.75 wt%) were obtained. Pyrolysation of the solid cake gave solid (34.5%wt), liquid (44.8 wt%) and gas (20.7 wt%) products. The obtained pyrolysed liquid had aqueous- and oil phase 66.7 wt% and 33.3 wt%, respectively. The reaction of the aqueous phase over the zirconia-iron oxide under the optimum condition (T=320 _C, W/F = 2 h) produced the liquid containing methanol (0.34 M), acetone (0.41 M) and acetic acid (0.06 M) with the price of 2,153 B/ton of aqueous phase. The reaction of the oil phase over the zirconia-iron oxide under the optimum condition (T=320 _C, W/F = 6 h) produced the liquid containing acetone (0.06 M), phenol (0.08 M) and m-cresol (0.008 M) with the price of 4,117 B/ton of oil phase. The experimental results showed that the catalyst deactivated during the reaction. However, the catalyst regenerated under air at 550 _C showed the reaction performance similar to the fresh one. To evaluate the performance of the overall process, the reaction of each phase was tested using a pilot scale reactor (capacity = 0.5 Lliquid waste/h). The large pilot scale reactor showed the recovery performance worse than the small experimental reactor. The poorer performance resulted from non-isothermal inside the large pilot scale reactor. The economical analysis based on 0.5 Lliquid waste/h scale showed that the proposed new process could be profitably operated: capital cost = 8.90 MB and profit = 8.88 MB/year.