Anaerobic digestion (AD) of sugarcane filter cake (SFC) was investigated by comparing the performance of pre-treatment methods in biochemical methane potential (BMP) tests and semi-continuous experiments. For that, SFC was pre-treated by autoclaving the substrate alone or with sodium hydroxide (NaOH). Experimental data from BMP tests were fitted to a kinetics model and further used for simulating the AD process in a continuous stirred-tank reactor (CSTR). BMP tests showed differences (p < 0.05) in total methane potential (SBMP), which have affected methane yields during simulation in a CSTR. Untreated produced 185 mL CH4 gVS−1, autoclaved pre-treatment 174 mL CH4 gVS−1 and autoclaved with NaOH pre-treatment 222 mL CH4 gVS−1. Interestingly, such higher performance of autoclaved with NaOH pre-treatment was only observed at earlier stages during semi-continuous feeding experiment. At steady-state no significant differences (p > 0.05) in terms of methane yield were observed among the reactors (average of 224 mL CH4 gVS−1). These results demonstrate that the benefits of pre-treatment could only be observed in BMP tests, which is likely explained by a better adaptation of the microbial community to the substrate during long term semi-continuous experiment, making SFC pre-treatment ineffective in a single-stage CSTR and under this feeding regime.