This work presents report of the detoxification of cassava wastewater using bamboo (Bambusoideae) leaf. The aim of the detoxification process is toreduce the poisonous cyanide content to a permissible level of using cassava effluent for biogas production. Fresh leaves of bamboo and cassava roots used for this experiment were obtained from Enugu, Nigeria. The effluent was obtained from the cassava roots through grinding and dewatering processes, while fine particles of the bamboo leaf were obtained through sun-drying, grinding and sieving processes. Fourier transform infrared (FTIR) spectroscopy aided the determination of the functional groups characteristic to the substances which are present in bamboo leaf. Physicochemical properties of the cassava effluent (pH, BOD, COD and cyanide content) were evaluated. BOD tests were carried out by measuring the amount of dissolved oxygen present in the samples before and after incubation at 20°C for 5 days. Alkaline picrate method was used to determine the concentration of the cyanide. The cyanide content reduction was optimized using response surface methodology. It involved the use of central composite design of Design Expert software. The considered factors were particle size, mass/volume ratio, temperature and time while the response was cyanide content reduction. Analysis of variance (ANOVA) and mathematical modelling of the percentage removal of cyanide content were carried out. From the analysis of the results, the major functional groups responsible for the cyanide content removal werefound to be: C–H deformation, C=O stretching, C–H stretching, C=C=C stretching and O–H stretching. The presence of bamboo leaf reduced the cyanide content to 0.12 mg/L, a permissible level that is harmless to enzymatic reaction of biogas production. Quadratic model adequately described the percentage removal of the cyanide content. Optimum cyanide content removal of 83.6% was obtained at particle size of 0.458 mm, mass/volume ratio of 0.042 g/mL, temperature of 331.4 K, and time of 11.1 h. The model was confirmed by various criteria; correlation coefficient R2 (0.9715), p value (<0.0001) and adequate precision (24.365).