The invention discloses a resource allocation method for a server system. The method comprises the steps of predicting a resource usage amount of an application process by utilizing an artificial neural network algorithm; when the resource usage amount of the application process is greater than an available virtual machine resource threshold of the application process, starting a virtual machine for the application process to use; and adjusting the available virtual machine resource threshold to be the sum of the available virtual machine resource threshold and a resource amount of the virtual machine.