Sweetness is one of the main drivers of consumer preference, and thus is given high priority in apple breeding programmes. Due to the complexity of sweetness evaluation, soluble solid content (SSC) is commonly used as an estimation of this trait. Nevertheless, it has been demonstrated that SSC and sweet taste are poorly correlated. Though individual sugar content may vary greatly between and within apple cultivars, no previous study has tried to investigate the relationship between the amount of individual sugars, or ratios of these, and apple sweetness. In this work, we quantified the major sugars (sucrose, glucose, fructose, xylose) and sorbitol and explored their influence on perceived sweetness in apple; we also related this to malic acid content, SSC and volatile compounds. Our data confirmed that the correlation between sweetness and SSC is weak. We found that sorbitol content correlates (similarly to SSC) with perceived sweetness better than any other single sugar or total sugar content. The single sugars show no differentiable importance in determining apple sweetness. Our predictive model based on partial least squares regression shows that after sorbitol and SSC, the most important contribution to apple sweetness is provided by several volatile compounds, mainly esters and farnesene.