This paper examines the formation of functional metropolitan regions within the City of Seoul, and addresses the problem of forecasting such trends based on use of the OD matrix and analysis of movement patterns with defining complex linkage systems. For such purposes, the daily movements of people in Seoul have been taken as an analytic indicator of functional linkages. The OD data were collected in 1977 as part of a KIST (Korean Institute of Science and Technique) project.In the first stage of the analysis, prediction models are built using stepwise multiple regression analysis, using 19 explanatory variables about socio-economic characteristics and 12 dependent variables concerning generated/pulled trips for each individual's journey purpose (Table 2). The number of trips for 1991 is subsequently derived from prediction models for each zone.The next stage is a trip distribution process. The author has used a BPR-type model in various gravity models, with the following general form:tijp=Tip·ujpfp(Dij)·Kijp/n∑k=1ukpfp(Dik)·Kikpwhere: p=trip purposes;tij=total trips from zone i to zone j;Ti=total trips generated in zone i;uj=total trips pulled by zone j;Kij=constant for total trips from zone i to zone j; andf(Dij)=αdijβ exp (γdij), where α·β·γ are parameters, and dij is the road (or time) distance between zone i and zone j.Using the above model, the new OD matrix for 1991 is derived. 1991 is the targeted completion year for the Seoul city master plan. At approximately this time, the city's urban growth should enter a period of stability.The final step in the analysis involves the application of factor analysis rotation procedures, concentrating on the major patterns of common factors, and ignoring all other influences, due to problems concerning the diversity of the flows, and the limitations of cartographic methods. The matrix for daily flows of persons in Seoul consists of 100×100 elements. The Varimax rotation method is used here to facilitate the interpretation of latent factors. As a result of this application, seven functional regions can be identified in 1977 and 1991 respectively.For this study, only factor loadings greater than ±0.5 (1977) and ±0.4 (1991), and factor scores greater than ±2.0 (both years) are adopted as relevant indices. Inquiry into the seven sub-systems is done in a series of maps (Figures 4∼10), by linking each group of factor loadings (destinations) to its respective set of common factor scores (origins). However, zones with less 1% of total generated/pulled trips are excluded from linking objectives.Factor 1, accounting for 29.8% (1977) and 26.7% (1991) of the total variance, represents the most dominant nodal sub-system centered in the heart of Seoul, including the CBD and its surrounding areas. Factor 2, explaining a further 22% (1977) and 48.1% (1991) of the total variance, represents the secondary dominant sub-system centered in the Dobong area. These two factors, accounting for 52.5% (1977) and 48.1% (1991) of the total variance, represent the dominant sub-systems in the person flows of Seoul.The other five factors, accounting for 29.3% (1977) and 31.1% (1991), are as follows: Yeongdeungpo-Hwagog area, Cheonho area, Gwanag-Gangnam area, Cheongyangli area, and Sinchon-Eunpyung area. All of these seven sub-systems partly overlap each other at the periphery of each functional region. This is a result of a hierarchical pattern of functional regions.