Abstract: This paper introduces a serics of piecewise multiple general orthogonal polynomials,presents and demonstrates its basic properties and applies them to parameter identification for parameter separable systems.Due to the use of approximation technique by piecewise multiple low order orthogonal polynomials,the method manifests the following advantages:the amount of computation is little,the accuracy of results is high,it can be recurrently calculated and the algorithms do not need the initial estimation of the estimated parameters.Two algorithms have been presented and successfully applied to parameter identification for bacteria growth dynamic model of fermentation process.