This paper presents an Amended Lyrebird Optimization Algorithm (ALOBH) for band selection in hyperspectral images (HSIs). The high dimensionality of HSI data leads to computational complexity and redundancy. To address this, and to improve processing speed, efficiency, and classification accuracy, ALOBH selects a prime subset of bands. ALOBH is a modified population-based metaheuristic algorithm that is inspired by lyrebirds’ courtship displays and sound mimicry. The algorithm operates in two phases: exploration, influenced by the lyrebird’s escape strategy, and exploitation, inspired by its hiding scheme. ALOBH was tested on the Indian Pines and PaviaU HSI datasets. Experimental results demonstrate that ALOBH is an effective band selection method showing excellent overall, average, and classification accuracy, as well as a high convergence rate. Compared to the next best method, ALOBH improved average accuracy by \(2.2\%\) for Indian Pines and \(0.89\%\) for PaviaU.