Classical survival regression models may provide misleading results when event of interest occurs due to more than one causes. In this paper, taking all possible causes for the occurrence of event into account, a Truncated Poisson Exponential survival proportional hazard model has been proposed. An extensive simulation study has been conducted to examine the performance of the proposed survival model in the absence and presence of covariates under different percentages of censoring. The simulation results reveal that estimators of the regression parameters are consistent and efficient. To illustrate the model, under–five child survival data extracted from Bangladesh Demographic and Health Survey 2014 have been used.