To devise a multivariate parametric model for short-term prediction of disability using the Expanded Disability Status Scale (EDSS) and multimodal sensory EP (mEP). A total of 221 multiple sclerosis (MS) patients who underwent repeated mEP and EDSS assessments at variable time intervals over a 20-year period were retrospectively analyzed. Published criteria were used to compute a cumulative score (mEPS) of abnormalities for each of 908 individual tests. Data of a statistically balanced sample of 58 patients were fed to a parametrical regression analysis using time-lagged EDSS and mEPS along with other clinical variables to estimate future EDSS scores at 1 year. Whole sample cross-sectional mEPS were moderately correlated with EDSS, whereas longitudinal mEPS were not. Using the regression model, lagged mEPS and lagged EDSS along with clinical variables provided better future EDSS estimates. The R (2) measure of fit was significant and 72% of EDSS estimates showed an error value of ±0.5. A parametrical regression model combining EDSS and mEPS accurately predicts short-term disability in MS patients and could be used to optimize decisions concerning treatment.