Milk lipolysis, i.e. the hydrolysis of milk triglycerides, is not routinely measured and its occurrence and genetic determinism can therefore be hardly studied. In order to overcome this limitation, in France the LIPOMEC project (ANR-19-CE21-0010) has developed an equation for the prediction of goat milk lipolysis from mid-infrared (MIR) milk spectra. This prediction lipolysis equation was applied to non-standardized MIR spectra of milk samples of Alpine goats with a high diversity of αs1 casein genotypes. Finally, 22,745 log-normalized predictions of 1,035 goats were retained. First, environmental factors affecting milk lipolysis were identified. Then, variance components were estimated using a single trait animal model with repeatability and including the same environmental effects. The part of variance explained by the category of αs1 casein genotypes (for their known effect on protein content) was evaluated by including or excluding it in the model. Among the environmental factors identified, longer intervals between milking and milk analysis increased lipolysis in milk, as well as the evening milking compared with the morning milking. Lipolysis decreased with lactation stage, and its level was higher in the second lactation than in subsequent lactations. The lipolysis level was related to the category of αs1 casein genotypes, and reached a maximum for OO genotypes. Heritability of lipolysis was estimated to 0.21±0.03 and decreased to 0.18±0.03 when including αs1 casein genotypes in the model. The αs1 casein genotype would explain 3% of the total phenotypic variance of lipolysis. These first results will need to be verified using data from commercial farms in order to overcome the small size and the particular genetic structure of the population studied.