Introduction: Birth weight is a key predictor for risk of childhood illnesses and chances of survival; however in developing countries less than half of newborns are weighed at birth. In Nepal, only 36% of children born were weighed at birth. Nearly two thirds (63%) of deliveries take place at home and birth weight may not be known for many babies, the mother’s estimate of the baby’s size at birth could be used as an alternative. Aim and Objective: This study assessed the accuracy of low birth weight as perceived by mothers and factors influencing whether their perceptions were accurate. Methods: The study wasa facility based descriptive study carried out in four hospitals with sample size of 1533. Hospital nurses interviewed mothers using a pre-tested tool. Data was entered into EpiData 3.1 and analyzed using SPSS version 17 software package. Results: A total of 1533 mothers were interviewed of which 75 did not respond. An overall 75% mothers accurately identified actual low birth weight; and 25% mother perceived normal for actual low birth weight. Less percent of mothers <20years (sensitivity=0.74), illiterate (sensitivity=0.74), and primigravida (sensitivity=0.74) identified actual low birth weight than mothers ≥20years (sensitivity=0.75), literate (sensitivity=0.75) and multigravida (sensitivity=0.77). Conclusion: The study concluded that 75% mothers recognized actual low birth weight of newborn, and 25% mother’s perceived normal for actually low birth weight. The percentage of women accurately identifying actual low birth weight was slightly lower among mothers <20years, illiterate and primigravida as compared to mothers ≥20years, literate and multigravida.