If elections were held today, President Uhuru Kenyatta would get 53 per cent of the vote, according to a new opinion poll released Wednesday.
His closest challenger, NASA presidential candidate Raila Odinga would come second with 43 per cent. Three per cent of voters are still undecided, it said.
The survey by Centre for Africa Progress was conducted between July 28 and August 1, 2017.
According to the poll, President Kenyatta has strong support in Central, North Eastern, Eastern and Rift Valley regions while Mr Raila trounces him in Nyanza and Western.
The firm's chairman Abel Oyieko, while releasing the results at Stanley Hotel in Nairobi Wednesday, said the study was meant to establish Kenyans' voting patterns and their favourite presidential candidates.
- READ MORE
- Why schools, colleges are still facing uncertainty
- Raila Jnr triggered what we need from patriotic youth
- BBI will create equal chances for all: Raila
- Unless BBI fails, Raila may not be in 2022 ballot
The poll indicated 61 per cent of voters believed the Independent Electoral and Boundaries Commission (IEBC) would deliver a free and fair election, while 37 per cent thought otherwise. Three per cent were not sure.
Some 72 per cent of respondents said they would vote for candidates nominated by the party they supported, while 32 per cent would not and two per cent were undecided.
A majority of respondents, 58 per cent, felt Uhuru was the best presidential candidate when it came to promoting peaceful elections, while 40 per cent said it was Raila.
The poll indicated 40 per cent of voters believed Uhuru would deliver on his campaign promises if re-elected, while 35 per cent believed Raila would keep his promises. On the other hand, 18 per cent did not trust any of them, while seven per cent were undecided.
Mr Oyieko said: "Nairobi County leads in the number of registered voters followed by Kiambu and Nakuru while Samburu, Isiolo and Lamu have the least number of registered voters." There are 19,611,423 registered voters.
Some 9,995 voters across all 47 counties were interviewed. The target sample was 10,000 registered voters and the data had a one per cent margin of error.
The sample was selected using the random, multi-stage stratified method and proportionate population size.
Oyieko added that urban to rural migration was likely to affect voter turnout.