Many environmental problems result from human activities and 

are subjects of increasing concern in the world. 

Thus environmental improvement policies (EIPs) have become 

an issue of increasing importance. In this paper we consider 

a problem in which an agent implements the environmental 

improvement policy under uncertainty. If an emission level 

of a pollutant arrives at a critical level, the agent has to 

decrease the emission to a certain level in order to improve 

environment. The agent problem is to minimize the expected 

total discounted cost which includes a cost to implement the EIP 

and an associated damage from the pollutant under the assumption 

that a state process of the pollutant follows a geometric 

Brownian motion. Then we find critical emission levels of 

pollutant, an optimal implementation times, optimal sizes of 

implementation and evaluate the optimal EIP (OEIP) by using 

an impulse control method. Some numerical examples are practiced 

to illustrate our results.