A mechanistic model predicting the accumulation of tomato fruit sugars was developed in order (i) to dissect the relative influence of three underlying processes: assimilate supply (S), metabolic transformation of sugars into other compounds (M), and dilution by water uptake (D); and (ii) to estimate the genetic variability of S, M, and D. The latter was estimated in a population of 20 introgression lines derived from the introgression of a wild tomato species (Solanum chmielewskii) into S. lycopersicum, grown under two contrasted fruit load conditions. Low load systematically decreased D in the whole population, while S and M were targets of genotype×fruit load interactions. The sugar concentration positively correlated to S and D when the variation was due to genetic introgressions, while it positively correlated to S and M when the variation was due to changes in fruit load. Co-localizations between quantitative trait loci (QTLs) for sugar concentration and QTLs for S, M, and D allowed hypotheses to be proposed on the processes putatively involved at the QTLs. Among the five QTLs for sugar concentration, four co-localized with QTLs for S, M, and D with similar allele effects. Moreover, the processes underlying QTLs for sugar accumulation changed according to the fruit load condition. Finally, for some genotypes, the processes underlying sugar concentration compensated in such a way that they did not modify the sugar concentration. By uncoupling genetic from physiological relationships between processes, these results provide new insights into further understanding of tomato fruit sugar accumulation.