Multi-band Orthogonal Frequency Division multiplexing based Ultra Wide-band (MB-OFDM UWB) technology is one of the strong alternatives for high data rate wireless personal area networks (WPANs) with low power consumption. The capacity of such systems is degraded by multi-path fading, shadowing, multi-user interference and noise. To improve system capacity under these adverse effects, in this paper, we devise cross-layer time scheduling methods, Proportional Time Scheduling with Modiano Algorithm (PTS-MA) and Proportional Time Scheduling with Channel State Information (PTS-CSI), in which scheduling and link adaptation are performed using instantaneous bit error probability (IBEP) estimates obtained through Modiano's algorithm and our novel estimation technique, respectively. We evaluate the performance of the PTS schemes by using numerical experiments. Simulation results suggest PTS-CSI scheduler as the most promising candidate for practical MB-OFDM UWB WPANs with high capacity and fair throughput distribution.