Species potential habitats predicted via various techniques, e.g. MaxEnt modelling in case of the current research, provide helpful information in terms of conservation and management programs, prioritization of limited resources and relative decision makings. Previous chapter was concerned with the modelling of the distribution of the Persian leopard potential habitats across the entire country in a regional context. Aside from the evaluation techniques to assess the modelling procedures which were done in the last chapter, validating the modelling outcomes according to the field data is essential. Thus, this chapter is dedicated to the ground validation of the predictive maps in selected study areas to ensure the accuracy for further conservation and management activities. For this purpose, three provinces in northeast (region 1), northwest (region 4) and south (region 3) of Iran with different environmental characteristics are selected to conduct camera trapping, field visits and indirect sign surveys, obtaining expert and local people knowledge via questionnaire surveys, group discussions and interviews. Three threshold methods including equal training sensitivity and specificity (A), maximum training sensitivity plus specificity (B) and minimum training presence (C) were selected for the purpose of binary classification of the predictive maps developed earlier using the MaxEnt software. The results indicated more accuracy of the sensitivity and specificity based threshold rules rather than the minimum training presence. Yet, intersection of the validated binary maps leads to the final conclusion of the habitat suitability rate of 0.3 on the predictive maps as a value to safely identify the actual potential habitats where importance for leopard conservation planning is confirmed.