In this study, a two-step analysis has been done by using the annual data of Turkey's 1960-2013 periods. First, alternative credit-to-GDP gap calculations were made by using Hodrick-Prescott, Band-Pass filters, and Unobserved Components Model and their performances were compared. Reached findings showed that the credit-to-GDP gap estimations based on the Unobserved Components Model is more successful compared to others. In the second step, banking crisis experienced in Turkey in the studied period has been analysed based on Logit Model. Systemic crises and all banking crises defined as a binary variable were modelled separately. In these models, credit-to-GDP gap estimated in the first step were used as explanatory variable. The results show that the growth in the credit-to-GDP gap increases the probability of occurrence of both systemic and non-systemic banking crises.