Descriptive and predictive models of land use are potentially an important tool in aiding land use decision making, providing that the reliability of their results including model output uncertainty is accounted for, explained, and properly communicated to model users. This study aims to contribute to spatial analysis of land use through an integrated, spatially-explicit approach to uncertainty and sensitivity analysis (iU-SA) for temporal and non-temporal land-use models employed in evaluation and simulation tasks. Two land-use models representative of land management practices and the transition of natural landscapes to urban land use have been tested in conjunction with the proposed iU-SA methodology. More explicitly, this research follows a three-step procedure to investigate the applicability of iU-SA for (1) land-use evaluation models, (2) land-use forecasting models, and (3) methods of visualizing the result of iU-SA. In the first part, a methodology for local multi-criteria evaluation (LMCE) extended with iU-SA has been developed to focus on the effects of local criteria weights and modeling scale on the variability of model output. The results show that LMCE provides more heterogeneity compared to global approach, however, raising the question of scale dependency. This question has been addressed in a scale sensitivity analysis, which showed that both scale and criteria weights play an important role in the results and in the ranking of the final decision of the decision-making process. In the second part, using the example of an urban growth model, a meta-modeling approach has been developed for spatially explicit U-SA. The quality of results obtained with the meta- modeling approach is comparable to the full-order modeling approach, which is computationally costly. The study shows that the meta-modeling approach can significantly reduce the computational effort of carrying out spatially explicit iU-SA in the application of spatio-temporal models. The overarching rationale of this dissertation has been to develop methods for better decision making. Therefore, in the final step, the effectiveness of the different visualization techniques for SA outputs is examined as a communication tool for decision making. The visualization of SA and user confidence dependency has been examined based on an empirical study of a web-designed survey applied for participants coming from different levels of expertise. Although the results did not show a statistically significant difference between compared visual representations, the dominant maps representation improved the correct answer rate for the expert users. The findings are insightful for the future works on improving the proposed techniques for SA visualization.