Summary: Shift scheduling problem (SSP) is a complex NP-hard integer programming problem, especially when many shifts and large number of workers of various ranks and multi-skills are involved. Shift scheduling also needs to comply with certain labour regulations and organisation’s rules and policies. Various models for SSP have been developed and solved. However, studies on SSP for workers of power stations or utility providers are still lacking. This paper presents the study on shift scheduling of workers at the largest power station in Malaysia. The objectives of this study are to identify main criteria and conditions of SSP at the power station, to formulate a binary integer goal programming (BGP) model for SSP that optimises workers’ day-off preference and to determine the optimal schedule for workers based on the model. The scheduling focuses on three processes which are demand modelling, shift scheduling and day-off scheduling. The study involves scheduling 43 workers in a selected department of the power station for 28 days where workers work in three shifts (morning, evening and night shifts) and have standby and rest days. The SSP-BGP model considers workers’ day-off preference as its main feature and introduces objective function that maximises the overachievement and minimises under-achievement of the day-off preference. It addresses conflicting multi-objectives in determining the optimal schedule using the goal programming approach. The model also put forward new hard and soft constraints, which reflect the nature of shifts’ requirements for these workers when determining the optimal schedule. The required data have been obtained from the power station and validation and verification of the model has been acquired with the cooperation from its personnel. The SSP-BGP model was solved using MATLAB in reasonable time. It significantly reduces the time to obtain a monthly schedule as compared to current manually done scheduling. Based on the 28-day schedule obtained, the day-off preference’s satisfaction of workers has increased by 37.21% that is from 43.02% based on the existing prepared schedule to 80.23% according to the schedule produced based on the SSP-BGP model. The model has shown good performance by not only generating optimal schedule based on the required number and composition of workers for shifts but also providing positive impacts on workers through day-off preferences’ satisfaction and the possibility of working with different groups.