Adipogenesis, the differentiation process of adipocyte (fat cell) formation from precursor cells contributes to increase of fat tissue in obesity. We construct deterministic model of the transcriptional network of adipogenesis to include a module for adiponectin (AdipoQ) production, an insulin-sensitizing hormone secreted by adipocytes, and its transcriptional regulators PPARgamma, and C/EBPalpha. Previous work has found that adiponectin levels increase and then decrease significantly during differentiation. Three coupled network modules found in adipocytes can explain these observations. PPAR gamma and C/EBP alpha trigger an incoherent feedforward loop in adiponectin production, leading to a pulse of adiponectin that disappears after a few days. Menawhile, a coherent feedforward loop in lipid synthesis leads to a delayed production of lipid droplets. We analyze two possible implementations for the adiponectin module to determine if variability within the system parameters alone is sufficient to explain the observed heterogeneity in adipocytes. Our results show that only the model where lipid droplets increases the degradation of adiponectin fits the trends observed in experimental studies, indicating that it is more likely than the model where lipid droplets inhibit the synthesis of adiponectin.