A 2-yr grazing experiment was conducted to assess the effects of grazing management on cattle distribution and pasture and stream bank characteristics. Six 12.1-ha cool-season grass pastures in central Iowa were allotted to 1 of 3 treatments: continuous stocking with unrestricted stream access (CSU), continuous stocking with stream access restricted to 4.9-m-wide stabilized crossings (CSR), or rotational stocking with stream access restricted to a riparian paddock (RP). Pastures were stocked with 15 fall-calving Angus cows (Bos taurus L.) from mid-May to mid-October for 153 d in 2008 and 2009. A global positioning system (GPS) collar recording cow position every 10 min was placed on at least 1 cow per pasture for 2 wk of each month from May through September. Off-stream water was provided to cattle in CSU and CSR treatments during the second of the 2 wk when GPS collars were on the cattle. A black globe temperature relative humidity index (BGTHI) was measured at 10-min intervals to match the time of the GPS measurements. Each month of the grazing season, forage characteristics (sward height, forage mass, and CP, IVDMD, and P concentrations) and bare and fecal-covered ground were measured. Stream bank erosion susceptibility was visually scored in May, August, and October (pre-, mid-, and post-stocking). Cattle in RP and CSR treatments spent less time (P < 0.10) within the stream zone (0 to 3 m from stream center) in June and August and in the streamside zone (0 to 33 m from stream zone) in May through August and May through September, respectively, than cattle in CSU pastures. However, off-stream water had no effect on cattle distribution. Compared with the CSU treatment, the CSR treatment reduced the probability (P < 0.10) that cattle were within the riparian zone (0 to 36 m from stream center) at BGTHI of 50 to 100. Bare ground was greater (P < 0.10) in pastures with the CSU than CSR and RP treatments in the stream and streamside zones in September and October and in July and September. Streams in pastures with the CSU treatment had less stable banks (P < 0.10) mid- and post-stocking than RP or CSR treatments. Results show that time spent by cattle near pasture streams can be reduced by RP or CSR treatments, thereby decreasing risks of sediment and nutrient loading of pasture streams even during periods of increased BGTHI.