This thesis develops a model aimed at predicting the performance of bus crew schedules and study their operational capacity to face disruptions, that is, its robustness. It is based on the relation among schedule deviation, passengers and slack times, considered as non-productive times before trips.Optimization of Public Transport is a challenge that main cities are facing since this type of transport is the backbone of mobility. New powerful versions of transport planning software are providing more optimal schedules to operators, therefore decreasing their costs, but increasing the risk of delays due to more gathered schedules that can cause fines and passengers’ dissatisfaction.A trade-off between production and fines costs is needed to find the optimal slack times. To reach that goal, the Frihamnen bus depot in Stockholm (Sweden) was selected as case study, having historical data of all their lines for last year. A regression model for schedule deviation was calculated which served to create a computer tool in Microsoft Excel, giving the possibility to checking the performance of new schedules based on that historical data.The results showed an inverse relation between schedule deviation and slack times. Moreover, short non-productive times before trips also mean high delays on departures. A positive output is that adding a stop of at least 10% of the total trip time, the first departure would be on time or 2 minutes late as maximum. Several software developers are applying similar studies to their products to keep optimizing bus schedules while taking an economic and social approach.