We analysed to what extent the stochastic effects of two biological control agents (i.e. weevils and mycoherbicides) affect the optimal choice of Californian thistle control. A stochastic, dynamic optimisation model was set up to analyse strategies that maximise the expected net present values. We analysed the cost-effective strategies to control the thistle for deterministic and stochastic cases. Results show that the stochasticity of the efficacy of weevils does not affect the optimal strategy. Compared to the deterministic case, however, mycoherbicides will be introduced at a higher level of weed density if we take the stochastic effect of mycoherbicides into account.