SUMMARY
To develop community-based early warning systemof forest and land fires is needed local knowledges exploration that have correspondence to existing systems. The objectives of this study were to determine characteristics offorest and land fires and to explore local knowledge related to forest and land fire early warning system. Spatial analysis with geographic information system of hotspot data,rainfall spatial distribution and fire driving factors were used to determine temporal and spatial characteristics of forest and land fires. Local knowledge regarding of dry season prediction, fire location and causes of fire were determined by depth interviews and focus group discussions. Results of this study showed that temporal and spatial analysis of hotspot data, rainfall and fire driving factors have correspondency to local knowledge offorest and land fires early warning. Hotspot and rainfallanalysis also correspondend tolocal knowledge about prediction of risk and monitoring of forest and land fires. Spatial analysis of hotspot data with fire driving factorsnamely land cover, river, road and land types also correspondend to public knowledge about monitoring, analysis and prediction of hazard. We conclude that local knowledge have big potency that strengthen institutionally in controlling forest fires and land in Kapuas.