Cassava (Manihot esculenta Crantz) like most other crops responds to the effect of genotype by environment interaction (GEI), which makes it difficult to identify the superior genotypes in terms of performance, stability and adaptability. Currently there is limited information about the adaptability of cassava genotypes at different times of harvesting (TOH). Due to increasing demand for early storage root bulking varieties, and confounding effects of site, crop age, and season during selection, there is a need for the objective characterization of genotypes in terms of adaptability and stability with respect to TOH. The study was, therefore, conducted to identify high yielding, stable and adaptable cassava genotypes at different times of harvesting (6, 9 and 12 months after planting, MAP) through the application of multivariate analysis techniques. The study was conducted in four environments using sixteen genotypes using a triple lattice design. Variance components for individual environment were analysed using Restricted Maximum Likelihood (REML) while combined analysis was performed using Additive Main effects and Multiplicative Interaction (AMMI) model. The AMMI analysis of variance at three TOH revealed that variances due to genotypes, environments, and GEI were significant for most of the traits. However, at 6 MAP, GEI was not significant for most of the traits. Significance of main effects indicated stability of some genotypes across environments while GEI significance indicated that some genotypes were specifically adapted to certain environments. The non-significance of GEI at 6 MAP for almost all traits means that genotypes can be reliably evaluated in any single environment. Selection of a) high yielding genotypes based on mean performance, b) stable genotypes based on GSI (mean performance and ASV) and c) adaptability of genotypes based IPCA1 versus IPCA2 identified five genotypes (Mulola, Phoso, Maunjili, Beatrice and Unknown) that exhibited consistent performance, stability and adaptability across the three harvest periods. These genotypes are therefore best candidates for production in any of the studied and other similar environments.