Udio vlage mrtvog sitnog šumskog goriva je središnja komponenta gotovo svih modela za procjenu opasnosti od šumskih požara te modela za predikciju ponašanja šumskih požara. Glavni cilj ovoga rada bio je izraditi operativno primjenjive modele procjene udjela vlage otpalih iglica alepskog bora (Pinus halepensis Mill.) i dalmatinskog crnog bora (Pinus nigra Arnold subsp. dalmatica (Vis.) Franco), u uvjetima ispod točke zasićenosti vlakanaca. Ravnotežni udio vlage i vrijeme reakcije eksperimentalno su određeni u komori rasta, za obje vrste. Temeljem rezultata navedenih laboratorijskih eksperimenata i strukture HFFMC (Hourly Fine Fuel Moisture Code) modela, izrađeni su modeli za procjenu udjela vlage sitnog goriva na bazi sata, koji su u odabranim sastojinama navedenih vrsta i testirani, zajedno s izvornim HFFMC modelom. Rezultati testiranja su pokazali da HFFMC model, iako namijenjen za listinac bora, nije prikladan za operativnu primjenu u šumama alepskog i dalmatinskog crnog bora. S druge strane, izrađeni modeli su postigli bolje rezultate i prikladni su za operativnu primjenu jer srednja apsolutna pogreška modela za alepski bor iznosi 1 % udjela vlage, a modela za dalmatinski crni bor 0,86 % udjela vlage. Dobiveni rezultati sugeriraju da ovi modeli moraju biti specifični za pojedinu vrstu kako bi postigli točnost potrebnu za preciznu predikciju ponašanja šumskog požara. U ovom su radu prvi put, za obje vrste, određeni ravnotežni udio vlage i vrijeme reakcije, ali i izrađen ovakav tip modela. Ravnotežni udio vlage prvi put je određen i za lišće hrasta crnike (Quercus ilex L.) te hrasta medunca (Quercus pubescens Willd.). Osim izrađenih modela za procjenu udjela vlage sitnog goriva, u radu je regresijskom analizom definirana povezanost debljine šumske prostirke s njenom količinom. Za sastojine alepskog bora, dalmatinskog crnog bora, hrasta crnike i hrasta medunca dani su regresijski modeli za OL-podhorizont, zatim za OF i OH podhorizonte zajedno, te za ukupnu količinu šumske prostirke. Izrađeni regresijski modeli omogućuju jednostavno kvantificiranje količine goriva, što je jako važno u operativnoj primjeni modela za predikciju ponašanja šumskih požara. Regresijski modeli, zajedno s izrađenim modelima za procjenu udjela vlage sitnog goriva, predstavljaju temelj za usvajanje postojećih ili izradu vlastitih modela za predikciju ponašanja šumskih požara. The moisture content of dead fine forest fuels is a central component of nearly all fire behaviour and fire danger rating systems. The main objective of this thesis was to develop operationally applicable fine fuel moisture prediction models for Aleppo pine (Pinus halepensis Mill.) and Dalmatian black pine (Pinus nigra (Arnold) subsp. dalmatica ((Vis.) Franco)) needle litters below the fibre saturation point. Experimental measurements of equilibrium moisture content and response time of dead needles were carried out in a growth chamber for both species. The results of these laboratory experiments and the structure of the HFFMC (Hourly Fine Fuel Moisture Code) model were used to develop hourly fine fuel moisture prediction models, whose predictive ability was tested in selected pine species stands. The predictive ability of the HFFMC model for both litter types was also evaluated. Although the HFFMC model was designed for pine litter, the results suggest that it is unsuitable for operational use in Aleppo pine and Dalmatian black pine forests. In contrast, new models performed better and can be used operationally because the mean absolute error of fine fuel moisture prediction model for Aleppo pine is 1% moisture content, and for Dalmatian black pine it is 0.86 % moisture content. The results suggest that these models should be species-specific in order to achieve the accuracy needed for precise prediction of fire behaviour. This thesis is the first to report equilibrium moisture content and response time data for both species, as well as develop this type of model. Equilibrium moisture content was also determined for the first time for the leaves of holm oak (Quercus ilex L.) and pubescent oak (Quercus pubescens Willd.). The other objective of this thesis, in addition to developing fine fuel moisture prediction models, was to develop regression models that relate forest floor depth to forest floor fuel load. Regression models were given for the OL horizon, the OF and OH horizons, and for the entire forest floor in Aleppo pine, Dalmatian black pine, holm oak and pubescent oak stands. These regression models allow for reliable quantifications of forest floor fuel loads, which is very important in the operative application of fire behaviour prediction models. Regression models, together with fine fuel moisture prediction models, provide a basis for the adoption of existing or development of our own fire behaviour prediction models.