In the present study, we assessed the degree to which various bull sperm traits, including the ability to bind to homologous zona pellucidae (ZP) and to fertilize in vitro, were related to field fertility after AI as 56d-non-return rates (56d-NRR based upon 8620 AIs) with frozen-thawed semen from 12 Swedish Red and White young (11-13 months) bulls (3 freezing batches/bull, 1-2 consecutive ejaculates/batch), entering a national breeding program and housed at two AI-bull stations. Significant (p < 0.01) correlations among in vitro tests were found for swim-up separated spermatozoa (concentration and concentration of motile spermatozoa) with the ZP-binding assay (ZBA, r = 0.72 and r = 0.73, respectively) and blastocyst rates (r = 0.79 and r = 0.79, respectively). A strong correlation was found between cleavage and blastocyst rates (r = 0.81, p < 0.01). Whereas bull station influenced (p < 0.05-0.001) all these variables, bull only affected (p < 0.05) ZBA and cleavage. There was no significant difference in 56d-NRRs (p > 0.05, range 62-67.2%) among bulls, once the absolute NRRs had been corrected for season, area and parity. Neither bull, batch nor station had a significant effect on field fertility (p < 0.05). None of the single sperm traits/tests in vitro correlated significantly with in vivo fertility. A correlation with bull fertility was, however, found when measures of sperm traits were combined and predicted NRRs, calculated by regression analyses, were tested against observed 56d-NRRs. The predictive bull fertility, based on the tested three frozen batches, varied from 61.8 to 67.5% (p < 0.05) and was highly correlated with the observed fertility NRRs (r = 0.92, p < 0.001). Although the results suggest that the combined analyses of swim-up separated sperm traits, ZBA and IVF can be used to predict bull fertility, it might not be possible to determine the life-time fertility of a breeding sire owing to the limited number of observations made.