Progressive censoring is essential for researchers in industry as a mean to remove subjects before the final termination point. Recently, kernel density estimation has been intensively investigated due to its nice properties and applications. In this paper we investigate the asymptotic properties of the kernel density estimators based on progressive type-II censoring and their application to hazard function estimation. A bias adjusted kernel density estimator is also suggested. Our simulation indicates that the kernel density estimates under progressive type-II censoring is competitive with kernel density estimates under simple random sampling.