A computer program, MSFIT, was developed for nonlinear fitting of release data from controlled release devices. The program is written in the C language, based on the high level user-interface Macintosh operating system. Five popular release models are implemented in the program. Initial estimates of the parameters are obtained by linear transformation followed by linear fitting of the fraction released-time data. Nonlinear model fitting is based on the maximum likelihood estimation method and is performed by the Levenberg-Marquardt method based on the χ 2 criterion. Several methods are available to aid the evaluation of the fitting results. The MSFIT program was used for the nonlinear fitting of several release data sets and results were satisfactory.