A commonly used approach to quantifying growth is to fit mathematical models to length-at-age data. Growth of the silver-lip pearl oysters, Pinctada maxima, cultured at a commercial pearl farm in West Papua, Indonesia was expressed mathematically by fitting five growth models (Gompertz, Richards, Logistic, Special von Bertalanffy Growth Function (VBGF) and General VBGF) to length-at-age data. The criteria used to determine the best fit model were a low mean residual sum of squares (MRSS), high coefficient of determination (r²) and low deviation of the asymptotic length (L_∞) from the maximum length (L_max). Using these criteria, the models were ranked accordingly: Special VBGF; General VBGF; Gompertz; Richards and Logistic models. The Special VBGF yielded the best fit (L_∞ = 168.38 mm; K = 0.930 year^–1; t₀ = 0.126; MRSS = 208.64; r² = 0.802; Deviation of L_∞ from L_max = 37.52 mm) and, accordingly, was used to model the growth of oysters cultured at three sites and two depths within the farm. Likelihood ratio tests were used to compare growth of oysters cultured at these sites and depths. Based on L_∞ and K values, favourable sites and depths could be determined that optimised growth requirements for the various stages of P. maxima culture. Sites with high K and L_∞ values were preferred sites for culturing juvenile oysters before pearl production, when high growth rate is essential to produce large numbers of oysters in the shortest time possible. In addition, high L_∞ may facilitate implantation of larger nuclei conducive to the production of larger, more valuable pearls. Conversely, sites with low K values were preferred sites for weakening P. maxima before pearl ‘seeding’, a process undertaken to minimise nucleus rejection after seeding.