Intracranial pressure (ICP) is a key vital sign whose temporal behavior across multiple time scales is in- creasingly being studied, e.g., to assess brain compliance or to estimate cerebral perfusion. In this paper we present a multiscale ICP signal simulator that acts on three time scales: Long-term ICP trends (minutes to hours), respiration-induced changes in ICP and single ICP pulses. The long-term trends are based on a mechanistic ICP model, whereas the respiration-induced and intrabeat changes are represented through phenomenological models. The latter capture waveform properties extracted from real patients in our ICP database as well as waveform features described in the literature. The presented ICP simulator could be used to evaluate ICP signal analysis algorithms and as a mock signal for calibrating and testing patient monitors.