The construction of a habitat suitability index (HSI) model is a crucial problem in fishing ground forecasts. In general, the HSI model is established by estimating the relationship between marine environmental factors and fishing ground probabilities. However, the environmental factors observed by remote sensing technology and commercial fishing investigations are usually highly correlated, and conventional methods such as the continued product model, minimum model, maximum model, arithmetic mean model, and geometric mean model cannot eliminate the harmful effects caused by the correlation of fishing data. As a result, it is difficult for them to capture the complex relations between environmental factors and fishing ground probabilities. Based on the widely used intelligent optimization method of genetic algorithms (GAs), this paper presents a general framework called GeneHSI for HSI modeling and intelligent optimization. Most importantly, the GeneHSI framework can remove the harmful effects of correlation, allowing the automatic retrieval and optimization of the HSI parameters. The core of GeneHSI modeling is the construction of a fitness function. This function was built by projecting the logistic regression-based HSI space to that of a GA, and is used to guide the optimization process of GeneHSI. Specifically, the fundamental concept of the projection is to minimize accumulative errors between the computed ground probabilities and the observed probabilities converted from commercial fishing data. The proposed GeneHSI framework is composed of three elements. These are the construction of the problem to be solved, the initialization of the GA, and the optimization strategy of the GA. The validation and effectiveness of the GeneHSI framework have been demonstrated using simulation data, that is, randomly generated normalized marine environmental factors and fishing ground probabilities range from 0 to 1. Research shows that the GeneHSI framework is effective and efficient in retrieving and optimizing HSI parameters for fishing ground forecasts. Because of the stochastic characteristics of GAs, however, there is a high requirement for modelers and scientists to better control the implementation of the GeneHSI framework. The HSI parameters retrieved by the GeneHSI framework vary under different constraints. Such constraints used in GAs commonly include linear inequalities and linear equality constraints on the underlying relations between marine environmental factors and fishing ground probabilities, as well as constraints on the bounds of HSI parameters. Compared with the results under optimization strategies using these constraints, the results under a general optimization strategy are inferior in that the GeneHSI framework cannot obtain a good match between the best-fitness and mean-fitness curves. In theory, the fitness value is the accumulative error of the GeneHSI model; hence, a smaller value indicates a better result. However, a good convergence process does not necessarily lead to a minimum fitness value amongst fitness functions under different constraints. In this paper, therefore, an evaluation of the convergence process, instead of a minimum fitness value, is considered the fundamental standard for the assessment of a good set of HSI parameters. In addition, experience and professional knowledge are required for an exact assessment of the HSI parameters. Overall, the above constraints, especially those on the parameter bounds, greatly help the optimization of the GeneHSI framework to retrieve better HSI parameters. In addition, the implementation of the GeneHSI framework with 100, 1 000, 5 000, and 10 000 samples demonstrates its strong capability for processing the mass data of fishing grounds. It is expected that the GeneHSI framework can enrich the modeling methods and theories of fishing grounds, and hence guide the application of intelligent optimization methods in fishing ground HSI modeling.