Potato vigour, the growth potential of seed potatoes, is a key agronomic trait that varies significantly across production fields due to factors such as genetic background and environmental conditions. Seed tuber microbiomes are thought to influence plant health and crop performance, yet the precise relationships between microbiome composition and potato vigour remain unclear. Here we conducted microbiome sequencing on seed tuber eyes and heel ends from 6 potato varieties grown in 240 fields. By using time-resolved drone imaging of three trial fields in the next season to track crop development, we were able to link microbiome composition with potato vigour. We used microbiome data at varying taxonomic resolutions to build random forest predictive models and found that amplicon sequence variants provided the highest predictive accuracy for potato vigour. The model revealed variety-specific relationships between the seed tuber microbiome and next season’s crop vigour in independent trial fields. With a coefficient of determination value of 0.69 for the best-performing variety, the model accurately predicted vigour in seed tubers from fields not previously included in the analysis. Moreover, the model identified key microbial indicators of vigour from which a Streptomyces, an Acinetobacter and a Cellvibrio amplicon sequence variant stood out as the most important contributors to the model’s accuracy. This study shows that seed potato vigour can be reliably predicted based on the microbiota associated with seed tuber eyes, potentially guiding future microbiome-informed breeding strategies.