The present work aims to improve a code for calculating k0 and Q0 parameters, called COVAR, adding another method of calculating k0 factor and improving the covariance analysis, creating a new version: COVAR v4.1. The present work also aims the development of a new method of calculating the alpha and several k0 parameters in a single least square fit, by means of a novel methodology, using covariance matrices and all partial uncertainties. For the calculations applying this new method, another code was developed, called AKFIT v2.1 which performs linear and non-linear fittings for the determination of alpha and k0 parameters for several irradiations in different periods. We used a database with irradiations in the years 2008 and 2010 performed at the IEAR-1 nuclear reactor of the IPEN-CNEN/SP, by the Nuclear Metrology Laboratory (LMN), corresponding to radionuclides 95 Zr, 65 Zn, 69m Zn, 46 Sc, 140 La and 60 Co and resulting in 21 data sets which were analyzed in order to verify the performance of COVAR4.1 and AKFIT2.1. For COVAR v4.1, the results with the alternative calculation of k0 factor were close to the already existing calculation and were consistent with the literature. For AKFIT v2.1 fittings were performed with both irradiations simultaneously and separately. The fitted models agreed with the literature. The value was 0,0025(83), which agrees with previous results obtained by the LMN. The correlations between the parameters k0 behaved as expected, with smaller values between different elements and greater correlations between equal elements with different energies and using the same comparator measurement. It can be concluded that the proposed methods were able to calculate the values of k0, with AKFIT v2.1 being a new technique in which it is possible to determine two parameters, alpha and k0 at the same time, quickly and accurately. It is expected that AKFIT code can be improved by adding more parameters, such as Q0 and f, by making a complete fitting for the determination of all the main parameters for the k0 NAA method.