Since non-motorized transportation has been excluded from the main stream of the general transportation planning, there is no standard technique for estimating non-motorized transportation demand. This thesis aims to develop a trip generation model for pedestrians. This research uses Poisson regression to model as walk trips can be regarded as event counts. The National Household Travel Survey (NHTS) 2001 is the primary source of data used, supplemented by MD Property View data. At the individual level, regression models for general walk trips and walk trips to work are estimated. Then, the walk trips to work are validated by Census data at the tract level. Finally, using that walk trips to work have similar tendency to total walk trips, this thesis estimates the number of total pedestrians a day for 506 tracts in Baltimore metropolitan region.