The primary objective of this study was to determine the accuracy of physiotherapists' predictions of mobility for people with spinal cord injury (SCI). Five physiotherapists and 47 patients from two SCI units were recruited. Two sets of predictions about mobility at 1 year were made for each patient shortly after they commenced rehabilitation. The predictions were made using standardized outcome measures. One set of predictions was made by the physiotherapists from the patients' SCI unit who had face-to-face contact with the patients and the other set of predictions was made by the physiotherapists from the other SCI unit who had no face-to-face contact with the patients and instead relied on chart records, and vice versa for patients from the second SCI unit. Correlations between mobility at 1 year and the predictions made by the physiotherapists who had face-to-face contact with the patients ranged between 0.54 and 0.85. The corresponding correlations for the physiotherapists who did not have face-to-face contact ranged between 0.40 and 0.71. Contact and non-contact physiotherapists' predictions were within one point of outcomes between 55% and 83% of the time. Physiotherapists were reasonably accurate at predicting future mobility. Face-to-face contact with patients slightly improved the accuracy of predictions.