Soil surveys, soil maps, spatial soil information systems, soil geographic databases, are all designed for fulfilling the requirements and demands of society, which increased dramatically in the last decades. Traditional soil survey is time consuming and expensive, new conventional surveys in the near future are very unlike, consequently methods exploiting existing information are becoming increasingly important. In the recent digital era spatial soil information systems (SSISs) are playing a more and more important role in this context. A key issue of applicability of SSISs is their accuracy. Essentially, the main practical aim of soil surveys and soil maps is prediction. It simply means that certain soil feature is estimated for a whole region based on available soil data collected at localized sample points. The traditional tool of this information extension is the classical soil map using soil mapping units. Crisp soil maps subdivide the region into disjunctive units in a way that within heterogeneity of soil properties is less than for the whole territory. Numerous novel methods have been developed for producing more accurate soil maps; traditional crisp soil maps however are still extensively applied, since they offer the most easily interpretable results for the majority of users. On the other hand accuracy of crisp soil maps can be increased in several ways: with the refinement of soil contours; with the subdivision of mapping units taking into consideration smaller within patch inhomogeneities; and with the refinement of attribute information (more recent data, more precise measurement, up-to-date methodology, more appropriate classification etc.). A great amount of soil information is available in Hungary due to former agrogeological surveys. The collected data are available in different scales: national, regional, micro-regional, farm and field level and generally they are related to maps. However, similarly to the great majority of the world, large scale, comprehensive new surveys cannot be expected in the immediate future. In the 1990s a great deal, dominantly the small-scale of these soil related data were converted into digital format and organized into SSISs. Recently the more challenging large-scale systems are coming to the front. As primary importance, GIS adaptation and digital reambulation of the results of the practical 1:25,000 scale soil-mapping programme hallmarked by Kreybig is under construction. There is much more utilizable information originating from this survey, than it was processed traditionally and published on the map series and in reports, and what is provided by simply archiving them digitally. A true SSIS can and should reach higher levels of digital processing. Integration of Kreybig Digital Soil Information System within appropriate spatial data infrastructure (SDI) and its updating with efficient field correlation make an inherent refinement and upgrading of the system possible as well as the estimation, measurement of the reliability of the system. As a result, the raw information processed using appropriate methods together with complementary spatial, digital, environmental data, a higher level, more accurate and consequently more reliable system could be developed. During its development different stages of KDSIS provides soil information on different level of accuracy. This kind of multilevel feature can be also preserved and even utilized. KDSIS provides two different opportunities for increasing its spatial and thematic accuracy on soil properties: (i) spatial refinement of spatial units integrating Kreybig profile methodology and SDI; (ii) updating soil information with recent fieldwork. In our paper we present GIS based methods developed for the spatial and thematic refinement, improvement of Kreybig Digital Soil Information System. Upgraded KDSIS makes the compilation of up-to-date (crisp) soil maps possible. Mapping units of the old and new maps may differ due to several reasons. There are reshaped contours, new soil patches, and units with changed attributes. Implication of new sampling data collected at revisited sites makes the comparison of archived (and so far stored) and newly surveyed data possible. Thus changes in soil properties can be identified. This in one hand should be recorded in the database thus updating it. On the other hand trends can be identified in soil characteristics and functions, degradation processes can be realized and/or forecasted. It can serve as reference to the study of anthropogenic effects. Joint management and application of multi-temporal spatial soil information within an appropriate relational database management system (RDBMS) and GIS environment makes KDSIS also a spatio-temporal soil information system.