Different sets of dry spell length such as complete series, monthly maximum, seasonal maximum, and annual maximum are applied and modeled with different probability distribution functions (such as Gumbel Max, generalized extreme value, Log-Logistic, generalized logistic, inverse Gaussian, Log-Pearson 3, generalized Pareto) to recognize in which duration, dry spells cause drought. The drought situation and temporal analysis in the North of Iraq region were done using the SPI index and by software of DrinC at a time scale of 3.6 and 12 months. Because of applicability, availability of data and the aim of the study, SPI is selected to analyze the dry spells in this study. Based on the maximum length of the available statistical period, the statistics for the years 1980 to 2019 were used from nine meteorological stations for analysis. The results of the study showed the severity of drought during the study period which related to dry spells. The results of this research confirm the variation of drought occurrence with varying degrees in different time and different dry spells condition in Iraq.