The implementation of process analytical technology(PAT) in a commercial active pharmaceutical ingredient(API) recrystallization production process was introduced in this paper,in which the near infrared(NIR) spectroscopy technology was used to monitor the concentration adjustment process under the requirements of good manufacturing practice(GMP).The hardware and software of the monitoring system were qualified according to the requirements of a complex computerized system.Two prediction models for API concentration and moisture were developed by partial least square(PLS) algorithm,which performance was evaluated by root mean square error of calibration(RMSEC),root mean square error of cross validation(RMSECV),root mean square error of prediction(RMSEP) and determination coefficient(R2),respectively.In order to ensure the performance of the models,the linearity,range,accuracy,precision(repeatability),specificity and robustness were verified again according to the analytical method verification requirements.Finally,the system performance was tested to confirm that the PAT system meets the requirements for commercial operation.PAT control of the concentration adjustment process could greatly shorten the concentration adjustment time,save the steam energy consumption and testing costs,reduce the deviation of production,increase the consistency of different batches,and improve the product quality.