Long-term metal exposure risk assessment for aquatic organism is a challenge because the chronic toxicity of chemical is not only determined by the amount of accumulated chemical but also affected by the ability of biological regulation or detoxification of biota. We quantified the arsenic (As) detoxification ability of tilapia and developed a biologically based growth toxicity modeling algorithm by integrating the process of detoxification and active regulations (i.e., the balance between accumulated dose, tissue damage and recovery, and the extent of induced toxic effect) for a life span ecological risk prediction. Results showed that detoxification rate (k (dex)) increased with increasing of waterborne As when the accumulated metal exceeded the internal threshold level of 19.1 μg g( - 1). The k (dex) values were comparable to or even higher than the rates of physiological loss and growth dilution in higher exposure conditions. Model predictions obtained from the proposed growth toxicity model were consistent with the measured growth data. The growth toxicity model was also used to illustrate the health condition and growth trajectories of tilapia from birth to natural death under different exposure scenarios. Results showed that temporal trends of health rates and growth trajectories of exposed fish in different treatments decreased with increasing time and waterborne As, revealing concentration-specific patterns. We suggested that the detoxification rate is critical and should be involved in the risk assessments framework. Our proposed modeling algorithm well characterizes the internal regulation activities and biological response of tilapia under long-term metal stresses.