Depending on dimensionality and constraints, you may have luck with Levenberg-Marquardt approach, lsqcurvefit and lsqnonlin functions in optimization toolbox.

Hi! Use the cftool for solving the problem. In the main screen choose your data and write your parameters (in numerical constrains) and left as a parameter your unknown and the toolbox will give you the value and the correlation coefficient. good luck! 

Thanks for the responses, but cftools is not going to work in straightforward manner, because now i have three variable which are changing continuously and the main reason is the functional form . How you will use the (b^c)/d on x axis when c is unknown?