Background This study aimed to investigate the putative role of the triglyceride‐glucose index (TyG index) computed as lnTG (mg/dl) × glucose (mg/dl)/2 and derived proxies as predictors of adiposity and weight loss changes after a low‐calorie diet (LCD) intervention. Methods A total of 744 adult participants from the multicentre DIOGenes intervention study were prescribed a LCD (800 kcal/day) during 8 weeks. Body composition and fat content at baseline and after 8 weeks were estimated by DEXA/BIA. A multivariate analysis approach was used to estimate the difference in ΔWeight1–2 (kg), ΔBMI1–2 (kg/m2) or ΔFat1–2 (%) between the basal value (point 1) and after 8 weeks following a LCD (point 2), respectively. The TyG index at baseline (TyG1), after following the LCD for 8 weeks (TyG2) or the TyG index differences between both time points (ΔTyG1–2) were analysed as predictors of weight and fat changes. Results TyG1 was associated with ΔWeight1–2 (kg) and ΔBMI1‐2 (kg/m2), with β = 0.812 (p = .017) and β = 0.265 (p = .018), respectively. Also, TyG2 values were inversely related to ΔFat1–2 (%), β = −1.473 (p = .015). Moreover, ΔTyG1–2 was associated with ΔWeight1–2 (kg) and ΔFat1–2 (%), β = 0.689 (p = .045) and β = 1.764 (p = .002), respectively. Furthermore, an association between TyG2 and resistance to fat loss was found (p = .015). Conclusion TyG1 index is a good predictor of weight loss induced by LCD. Moreover, TyG2 was closely related to resistance to fat loss, while ΔTyG1–2 values were positively associated with body fat changes. Therefore, TyG index and derived estimations could be used as markers of individualized responses to energy restriction and a surrogate of body composition outcomes in clinical/epidemiological settings in obesity conditions.