The invention relates to a path particle swarm optimization algorithm of a gene controlled nonlinear dynamical system. Ordering x(t)=(x1,x2, and the like,x14) <T> belonging to a set of R<14> to be a state variable, and an admissible set is a formula, wherein x1, x2, and the like, x5 respectively represent extracellular biomass, glycerol, 1,3-PD, acetic acid, and ethanol; x6, x7, and the like, x14 respectively represent intracellular glycerol, concentration of 3-HPA, concentration of 1,3-PD, mR, free regulation protein concentration R, mGDHt, concentration of glycerol dehydratase GDHt, mPDOR, and concentration of 1-3-PD oxidordeuctase PDOR. The optimization algorithm gives most possible path identification of a glycerol dissimilation reaction, and a new definition of extracellular relative error is introduced, and an obtained result is good in stability. The optimization algorithm adds a plurality of determining conditions, so the optimization algorithm fits actual problems of continuous fermentation, and operation efficiency and accuracy are improved.