This paper presents a mathematical model to evaluate the power output variations from photovoltaic systems (PV) in collaboration with an energy system planning in an urban area, and applied the model to a demonstrating energy system. The evaluating model is developed from per-floor area PV power output variations on the basis of the past 1 year of actual solar insolation. The evaluation of the power variations is represented in terms of a per-floor area standard deviation curve. In case of considering two alternatives, battery-less and battery-installed PV systems, a correlative coefficient curve is calculated, then a total standard deviation curve representing the total PV power variations can be formulated on the basis of a random variable (RV) model. The model is applied as a tool to evaluate the power variation in an energy system. The energy system planning demonstration approaches the optimal planning by minimizing total cost under CO2 emission reduction and the power variation constraints. The results show the impacts of the constraint of power variation that limit the installation of PV system. The constraint results the increase of the proportion of PV system with battery installation against PV system without battery installation. Consequently, the constraint also degrades the maximum potential of the entire system to reduce CO2 emission corresponding to the decrease of PV system. © 2009 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.