With the continuous development of the Internet industry, the function and technology of instant translation software can be continuously improved. This paper takes the machine translation model based on neural networks as the research basis and proposes a translation model suitable for the use of English instant translation software. Through word vector transformation, natural language is transformed into digitized symbols, and the n -gram model is used to improve the fluency of the processing. Using the neural probabilistic language model, the conditional probability of language processing is calculated. Introduce a recurrent neural network and optimize and adjust it based on the original model. Design the model framework for instant translation software for a series of operations, such as language signal processing. The case study revealed the factors that influenced users’ experience of using the software and their expectations of the instant translation software. The results show that the OR estimate of users’ satisfaction is 3.1655, and a one-unit increase in satisfaction will increase the advantage of instant translation software by 3.17 times. The error rate of translation software, as well as the learning cycle, have different degrees of influence on the advantage of the software. The period 2001-2022 was chosen as the study period, and the number of published patents for instant translation software research was examined. The overall trend during the study period showed a gradual increase. The number of patent publications increased by 40.9% to 62 in 2017, compared to the previous year. Instant translation software's future development trend is positive.
