The independent Zimbabwe Election Support Network's report said that the anomalies opened the way for "double voting and other rigging intentions".
In its research, the group found some 2 344 voters between the ages of 101 and 110 still on Zimbabwe's voting rolls, a dubious figure in a country where the average life expectancy is a mere 44 years.
The report also cited a lawmaker who found that more than 500 dead voters had all been given the same birth date - January 1 1901.
"An accurate, credible voters' register is a prerequisite for free and fair elections," the group said.
President Robert Mugabe, who has ruled the country since independence 30 years ago, has called for an election this year following a tumultuous power-sharing government with his long-time nemesis Morgan Tsvangirai, who became prime minister in 2009 under a unity deal.
The disputed 2008 election that led to the unlikely unity government was marred by violence, and critics have long accused Mugabe and his loyalists of vote-rigging.
The election network said that its researchers had found more than 185 000 cases of the same voters being listed in two or more voting constituencies during Zimbabwe's last vote.
Tsvangirai's party narrowly won the parliamentary vote in 2008 but he then boycotted a presidential run-off to protest violence against his supporters by Mugabe militants and loyalist police and soldiers.
Friday's report was based on face-to-face interviews with voters and a computer analysis of lists in 102 of the nation's 1 900 voting wards.
The government's chief voting administrator, Registrar General Tobaiwa Mudede, is seen as a Mugabe loyalist and his office is also the custodian of the nation's deaths register.
The group said he refused to release electronic copies of the whole voters' roll of 5.5 million registered voters, and hard copies were only made available intermittently for a fee across the country.
Some researchers also were threatened with violence and barred from carrying out field work in some areas by Mugabe militants, the group said.
The study ended up with maximum margin of error of plus or minus 4% at a 95% confidence level, the network said.
It found that electoral officials had not made efforts to update the lists by deleting the dead and eliminating duplications.
"It being a legal requirement that all deaths are reported and entered into a register of deaths, it is a simple task to ascertain which voters have died," the group said.
Even when some deaths were not officially registered, officials were obliged by the law to strike off names where there were sufficient grounds to believe the voter had died.
Dead people occupied their "pride of place in the voters' roll and presumably vote from their permanent resting places", it said.