Working memory with analogous stimulus plays an important role in the implementation of higher order cognitive function.Experiments indicate that the brain apply the distributed representation algorithm to encode and actively maintain the analogous stimulus during the delay period, such as working memory with spatial orientation and direction.To demonstrate the spatial working memory, Amari type line attractor model has been proposed, in which the bump attractors represent orientation or direction.Based on Amari type line attractor model, characteristics of spatial working memory, such as drift, accuracy, and capacity, have been elaborately investigated.However, Amari type line attractor model was challenged by the recent experiments in which the population firing rate encodes the analogous stimuli such as the value of options.Therefore, to actively maintain the value signal, one population of neurons should keep the persistent activity with a firing rate monotonically dependent on the value of options.In this study, we simulated the working memory of the value of options using a randomly connected spiking neuron network.