According to at least one embodiment, a failure prediction server includes an interface, a memory, and a processor. The interface is configured to communicate with equipment including a component that deteriorates over time. The memory is configured to store measurement data acquired from the equipment. The processor is configured to set a monitoring start condition for failure prediction in the corresponding equipment based on the measurement data acquired as initial data from the equipment, accumulate the measurement data acquired from the equipment in the memory after the monitoring start condition is satisfied, and predict a failure of the component based on the measurement data accumulated in the memory.