Mutual funds remain the most favoured investment instrument in the extensive domain of finance due to healthy and blooming returns in its asset group. In order to take investment decision and gain maximum benefit from the highly evolving universal financial market with minimum risk, forecasting Net Asset Value (NAV) of mutual funds is an absolute necessity. This paper presents a NAV predictor model which is designed using the non-iterative Random Vector Functional Link Network (RVFLN). Further, the parameters of the model have been calibrated using a hybrid chaotic meta-heuristic technique that includes Differential Evolution (DE) algorithm and Flower Pollination Algorithm (FPA) in its learning stage. With the application of direct link between input and output nodes and randomized weights in RVFLN, an attempt has been made to enhance the prediction capability of the model by integrating the natural evolution features of DE and the pollination process of FPA along with chaos theory. Validation of efficacy of the proposed optimization technique (CDFPA) is accomplished using ten standard benchmark functions by assessing the convergence pattern against popular and recent meta-heuristics. The proposed RVFLN-CDFPA NAV forecasting model has been applied on real time data sets of three popular Indian mutual funds to predict one day ahead NAV. Also, this research presents a comparison analysis to look into the consequences of various activation functions and popular chaotic maps on the performance of the suggested model. The potential of the novel integrated model has finally been compared with other state-of-the-art approaches. The proposed model exhibits an improved performance over other RVFLN based predictor models like RVFLN-CHFPA, RVFLN-DFPA, RVFLN-FPA, RVFLN-DE, RVFLN-PSO, RVFLN-AO, RVFLN-RSA, RVFLN-AOA and RVFLN-SCA. Also, the novel integrated framework exhibits an improved performance of 61.37%, 72.41%, 71.16% and 69.61% in RMSE over RVFLN, ELM, CPNN-FPA and LPNN-CHFPA, respectively, for DSP fund. For UTI, an improvement of 64.54%, 83.71%, 83.30% and 55.70% in RMSE and for SBI, an improvement of 81.35%, 88.16%, 83.55% and 82.80% in RMSE are reported over RVFLN, ELM, CPNN-FPA and LPNN-CHFPA, respectively, which clearly reveal the competency of the proposed framework over other experimented models.