Genetic modifications, such as gene knockout technique, have become mainstream in metabolic engineering to produce desired amount of targeted metabolites through reconstruction of the metabolic networks. The production, however, does not often achieve desirable outcome. To this end, in-silico methods have been applied to predict potential metabolic network response and optimise production. Previous methods working on relational modelling framework, such as OptKnock and OptGene, however, failed at handling its multivariable and multimodal functions optimization algorithms. This paper proposes hybridising bacterial foraging optimizationg algorithm (BFO) and dynamic flux balance analysis (DFBA) to overcome problems in OptKnock and OptGene with a nature-inspired algorithm and also to couple kinematic variables in the model to predict production of succinate in E.coli model. In-silico results showed that by knocking out genes identifed by BFODFBA, production rate of succinate is better as when compared to OptKnock and OptGene.