The aim of this paper is the needs to analyze the survival time for patients with Covide-19b who remains in the hospital until death, so it is necessary to study the survival times and estimate the reliability. The problem of finding the best distribution that fits the data is the key idea to analyze the data accurately. Consequently, the misspecifying of the distribution that fit the data leads to poor quality inference criteria of the phenomenon, also leads to unreliable reliability estimations. Many data of sciences areas are of different probability distributions depending on the nature of the phenomenon within the studied communities Some of the data are represented simple phenomena that cope with a unique probability distribution, and some of which are very complex and heterogeneous systems that force the researchers to use probability distributions fits the behavior of this random phenomenon. Many works in the field of reliability, failure and survival times, and the function of reliability (survival) follows some common distributions such as the Exponential distribution, Weibull distribution and other distributions. In this paper we introduced the survival function that follows an important distribution in survival modeling, that is called the Lindley distribution with two Parameters, taking into account two forms of this distribution, one of them we proposed based on different forms of the probability density function and finding the survival function for the distribution and compared to other distributions using several methods of estimation including the Maximum Likelihood Estimator (MLE), (percentiles estimators) by using Monte Carlo simulation experiments and comparing using the Integrated Mean Square Error (IMSE), (-2$ \ln L $) and AIC to achieve the best estimate of survival function among the distributions, as well as a real data analysis conducted for the survival times for patients with COVID-19 stay in hospital until death. The proposed distribution fitted the data very well in the Maximum Likelihood method compared with the other distribution.