There are some weaknesses of using solar PV system especially when there is issue of soiling on the surface of solar PV panel. The consequences for absence of this such study can cause unanticipated cost in the operation of solar PV panel. The objective of this project is to study the trend of soiling rate over different time period and its effect on the performance of solar PV panel in Malaysia and to develop a simple prediction model for cleaning interval of solar PV system in Malaysia. The study was conducted on real-time basis on a building's roof. Measurements of solar irradiance, voltage, current and the mass of dust collected were performed from both clean and dirty panels. It was discovered that the Monthly Test was significant with 4.53 of performance drop. Further analysis was conducted by running prediction model for cleaning interval. Intersection of graph plotting and fixed cleaning cost gives answer of cleaning interval that can be performed. It can be concluded that for every two and half month is the recommended time interval to perform regular cleaning to maximise electrical power generation by solar PV system in Malaysia. © 2018 The Authors, published by EDP Sciences.